@article{de_graaf_sound_2024,

title = {The sound of Parkinson's disease: A model of audible bradykinesia},

author = {D. Graaf and R. Araújo and M. Derksen and K. Zwinderman and N. M. Vries and J. IntHout and B. R. Bloem},

url = {https://www.sciencedirect.com/science/article/pii/S1353802024000154},

doi = {10.1016/j.parkreldis.2024.106003},

issn = {1353-8020},

year  = {2024},

date = {2024-03-01},

urldate = {2024-03-01},

journal = {Parkinsonism & Related Disorders},

volume = {120},

pages = {106003},

abstract = {Introduction 

Evaluation of bradykinesia is based on five motor tasks from the MDS-UPDRS. Visually scoring these motor tasks is subjective, resulting in significant interrater variability. Recent observations suggest that it may be easier to hear the characteristic features of bradykinesia, such as the decrement in sound intensity or force of repetitive movements. The objective is to evaluate whether audio signals derived during four MDS-UPDRS tasks can be used to detect and grade bradykinesia, using two machine learning models. 

Methods 

54 patients with Parkinson's disease and 28 healthy controls were filmed while executing the bradykinesia motor tasks. Several features were extracted from the audio signal, including number of taps, speed, sound intensity, decrement and freezes. For each motor task, two supervised machine learning models were trained, Logistic Regression (LR) and Support Vector Machine (SVM). 

Results 

Both classifiers were able to separate patients from controls reasonably well for the leg agility task, area under the receiver operating characteristic curve (AUC): 0.92 (95%CI: 0.78–0.99) for LR and 0.93 (0.81–1.00) for SVM. Also, models were able to differentiate less severe bradykinesia from severe bradykinesia, particularly for the pronation-supination motor task, with AUC: 0.90 (0.62–1.00) for LR and 0.82 (0.45–0.97) for SVM. 

Conclusion 

This audio-based approach discriminates PD from healthy controls with moderate-high accuracy and separated individuals with less severe bradykinesia from those with severe bradykinesia. Sound analysis may contribute to the identification and monitoring of bradykinesia.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{rao_pdf_2025,

title = {(PDF) A Plug-and-Play Method for Guided Multi-Contrast MRI Reconstruction Based on Content/Style Modeling},

author = {C. Rao and M. Osch and N. Pezzotti and J. Bresser and L. Beljaards and J. Meineke and E. Weerdt and H. Lu and M. Doneva and M. Staring},

url = {https://www.researchgate.net/publication/384245741_A_Plug-and-Play_Method_for_Guided_Multi-contrast_MRI_Reconstruction_based_on_ContentStyle_Modeling},

doi = {10.48550/arXiv.2409.13477},

year  = {2025},

date = {2025-08-01},

urldate = {2025-08-01},

journal = {ResearchGate},

abstract = {PDF textbar Since multiple MRI contrasts of the same anatomy contain redundant information, one contrast can be used as a prior for guiding the reconstruction... textbar Find, read and cite all the research you need on ResearchGate},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{beljaards_ai-based_2024,

title = {AI-based motion artifact severity estimation in undersampled MRI allowing for selection of appropriate reconstruction models},

author = {L. Beljaards and N. Pezzotti and C. Rao and M. Doneva and M. J. P. Osch and M. Staring},

url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/mp.16918},

doi = {10.1002/mp.16918},

issn = {2473-4209},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {Medical Physics},

volume = {51},

number = {5},

pages = {3555–3565},

abstract = {Background Magnetic Resonance acquisition is a time consuming process, making it susceptible to patient motion during scanning. Even motion in the order of a millimeter can introduce severe blurring and ghosting artifacts, potentially necessitating re-acquisition. Magnetic Resonance Imaging (MRI) can be accelerated by acquiring only a fraction of k-space, combined with advanced reconstruction techniques leveraging coil sensitivity profiles and prior knowledge. Artificial intelligence (AI)-based reconstruction techniques have recently been popularized, but generally assume an ideal setting without intra-scan motion. Purpose To retrospectively detect and quantify the severity of motion artifacts in undersampled MRI data. This may prove valuable as a safety mechanism for AI-based approaches, provide useful information to the reconstruction method, or prompt for re-acquisition while the patient is still in the scanner. Methods We developed a deep learning approach that detects and quantifies motion artifacts in undersampled brain MRI. We demonstrate that synthetically motion-corrupted data can be leveraged to train the convolutional neural network (CNN)-based motion artifact estimator, generalizing well to real-world data. Additionally, we leverage the motion artifact estimator by using it as a selector for a motion-robust reconstruction model in case a considerable amount of motion was detected, and a high data consistency model otherwise. Results Training and validation were performed on 4387 and 1304 synthetically motion-corrupted images and their uncorrupted counterparts, respectively. Testing was performed on undersampled in vivo motion-corrupted data from 28 volunteers, where our model distinguished head motion from motion-free scans with 91% and 96% accuracy when trained on synthetic and on real data, respectively. It predicted a manually defined quality label (‘Good’, ‘Medium’ or ‘Bad’ quality) correctly in 76% and 85% of the time when trained on synthetic and real data, respectively. When used as a selector it selected the appropriate reconstruction network 93% of the time, achieving near optimal SSIM values. Conclusions The proposed method quantified motion artifact severity in undersampled MRI data with high accuracy, enabling real-time motion artifact detection that can help improve the safety and quality of AI-based reconstructions.},

note = {_eprint: https://aapm.onlinelibrary.wiley.com/doi/pdf/10.1002/mp.16918},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{volleberg_deep_2025,

title = {Deep Learning-Derived Plaque Burden for Intracoronary Optical Coherence Tomography: An Intravascular Ultrasound-Based Validation Study},

author = {R. H. J. A. Volleberg and D. Shin and S. Saitta and R. A. Shlofmitz and E. Shlofmitz and A. Jeremias and R. G. A. Waerden and J. Thannhauser and N. Royen and Z. A. Ali},

doi = {10.1016/j.jcin.2025.07.021},

issn = {1876-7605},

year  = {2025},

date = {2025-10-01},

urldate = {2025-10-01},

journal = {JACC. Cardiovascular interventions},

volume = {18},

number = {19},

pages = {2432–2434},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{volleberg_artificial_2025,

title = {Artificial intelligence-based identification of thin-cap fibroatheromas and clinical outcomes: the PECTUS-AI study},

author = {R. H. J. A. Volleberg and T. J. Luttikholt and R. G. A. Waerden and P. Cancian and J. L. Zande and X. Gu and J. Q. Mol and T. Roleder and M. Prokop and C. I. Sánchez and B. Ginneken and I. Išgum and S. Saitta and J. Thannhauser and N. Royen},

url = {https://doi.org/10.1093/eurheartj/ehaf595},

doi = {10.1093/eurheartj/ehaf595},

issn = {0195-668X},

year  = {2025},

date = {2025-09-01},

urldate = {2025-09-01},

journal = {European Heart Journal},

pages = {ehaf595},

abstract = {Coronary thin-cap fibroatheromas (TCFA) are associated with adverse outcome, but identification of TCFA requires expertise and is highly time-demanding. This study evaluated the utility of artificial intelligence (AI) for TCFA identification in relation to clinical outcome.The PECTUS-AI study is a secondary analysis from the prospective observational PECTUS-obs study, in which 438 patients with myocardial infarction underwent optical coherence tomography (OCT) of all fractional flow reserve-negative non-culprit lesions (i.e. target lesions). OCT images were analyzed for the presence of TCFA by an independent core laboratory (CL-TCFA) and OCT-AID, a recently developed and validated AI segmentation algorithm (AI-TCFA). The primary outcome was defined as the composite of death from any cause, non-fatal myocardial infarction or unplanned revascularisation at 2 years (±30 days), excluding procedural and stent-related events.Among 414 patients, AI-TCFA and CL-TCFA were identified in 143 (34.5%) and 124 (30.0%) patients, respectively. AI-TCFA within the target lesion was significantly associated with the primary outcome [hazard ratio (HR) 1.99, 95% confidence interval (CI) 1.02–3.90},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{luttikholt_detection_2025,

title = {Detection of large areas of thin-cap fibroatheroma in a recurrent STEMI patient using a novel artificial intelligence algorithm: moving from 2D to 3D},

author = {T. J. Luttikholt and J. Thannhauser and N. Royen},

url = {https://doi.org/10.1093/eurheartj/ehaf189},

doi = {10.1093/eurheartj/ehaf189},

issn = {0195-668X},

year  = {2025},

date = {2025-07-01},

urldate = {2025-07-01},

journal = {European Heart Journal},

volume = {46},

number = {27},

pages = {2712},

abstract = {Optical coherence tomography (OCT) is a valuable imaging tool in percutaneous coronary intervention (PCI), recommended for stent guidance and evaluation.1 Moreover, OCT-imaging can visualize high-risk plaques, such as thin-cap fibroatheroma (TCFA), which have prognostic value.2,3 However, manual OCT-interpretation is time-consuming, subject to interobserver variability4 and, most importantly, assesses TCFA in a two-dimensional, single-frame manner. It is likely that prognosis depends on TCFA-extent rather than presence alone, similar to lipid burden in near-infrared spectroscopy.Our group developed OCT-AID, an artificial intelligence (AI)-algorithm for OCT-segmentation (Supplementary data online, Video S1) and plaque characterization.5 OCT-AID enables automated quantification of TCFA-area, as demonstrated in the present case.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{volleberg_comprehensive_2025,

title = {Comprehensive full-vessel segmentation and volumetric plaque quantification for intracoronary optical coherence tomography using deep learning},

author = {R. H. J. A. Volleberg and R. G. A. Waerden and T. J. Luttikholt and J. L. Zande and P. Cancian and X. Gu and J. Q. Mol and S. Quax and M. Prokop and C. I. Sánchez and B. Ginneken and I. Išgum and J. Thannhauser and S. Saitta and K. Nishimiya and T. Roleder and N. Royen},

url = {https://doi.org/10.1093/ehjdh/ztaf021},

doi = {10.1093/ehjdh/ztaf021},

issn = {2634-3916},

year  = {2025},

date = {2025-05-01},

urldate = {2025-05-01},

journal = {European Heart Journal - Digital Health},

volume = {6},

number = {3},

pages = {404–416},

abstract = {Intracoronary optical coherence tomography (OCT) provides detailed information on coronary lesions, but interpretation of OCT images is time-consuming and subject to interobserver variability. The aim of this study was to develop and validate a deep learning-based multiclass semantic segmentation algorithm for OCT (OCT-AID).A reference standard was obtained through manual multiclass annotation (guidewire artefact, lumen, side branch, intima, media, lipid plaque, calcified plaque, thrombus, plaque rupture, and background) of OCT images from a representative subset of pullbacks from the PECTUS-obs study. Pullbacks were randomly divided into a training and internal test set. An additional independent dataset was used for external testing. In total, 2808 frames were used for training and 218 for internal testing. The external test set comprised 392 frames. On the internal test set, the mean Dice score across nine classes was 0.659 overall and 0.757 on the true-positive frames, ranging from 0.281 to 0.989 per class. Substantial to almost perfect agreement was achieved for frame-wise identification of both lipid (κ=0.817, 95% CI 0.743–0.891) and calcified plaques (κ=0.795, 95% CI 0.703–0.887). For plaque quantification (e.g. lipid arc, calcium thickness), intraclass correlations of 0.664–0.884 were achieved. In the external test set, κ-values for lipid and calcified plaques were 0.720 (95% CI 0.640–0.800) and 0.851 (95% CI 0.794–0.908), respectively.The developed multiclass semantic segmentation method for intracoronary OCT images demonstrated promising capabilities for various classes, while having included difficult frames, such as those containing artefacts or destabilized plaques. This algorithm is an important step towards comprehensive and standardized OCT image interpretation.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{volleberg_optical_2025,

title = {Optical Coherence Tomography in Motion: Potential Cause for Artifacts},

author = {R. Volleberg and P. Cancian and N. Royen},

url = {https://www.sciencedirect.com/science/article/pii/S1936879824016960},

doi = {10.1016/j.jcin.2024.11.004},

issn = {1936-8798},

year  = {2025},

date = {2025-03-01},

urldate = {2025-03-01},

journal = {JACC: Cardiovascular Interventions},

volume = {18},

number = {5},

pages = {680–681},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{cancian_attenuation_2025,

title = {Attenuation artifact detection and severity classification in intracoronary OCT using mixed image representations},

author = {P. Cancian and S. Saitta and X. Gu and R. L. M. Herten and T. J. Luttikholt and J. Thannhauser and R. H. J. A. Volleberg and R. G. A. Waerden and J. L. Zande and C. I. Sánchez and B. Ginneken and N. Royen and I. Išgum},

url = {http://arxiv.org/abs/2503.05322},

doi = {10.48550/arXiv.2503.05322},

year  = {2025},

date = {2025-03-01},

urldate = {2025-03-01},

publisher = {arXiv},

abstract = {In intracoronary optical coherence tomography (OCT), blood residues and gas bubbles cause attenuation artifacts that can obscure critical vessel structures. The presence and severity of these artifacts may warrant re-acquisition, prolonging procedure time and increasing use of contrast agent. Accurate detection of these artifacts can guide targeted re-acquisition, reducing the amount of repeated scans needed to achieve diagnostically viable images. However, the highly heterogeneous appearance of these artifacts poses a challenge for the automated detection of the affected image regions. To enable automatic detection of the attenuation artifacts caused by blood residues and gas bubbles based on their severity, we propose a convolutional neural network that performs classification of the attenuation lines (A-lines) into three classes: no artifact, mild artifact and severe artifact. Our model extracts and merges features from OCT images in both Cartesian and polar coordinates, where each column of the image represents an A-line. Our method detects the presence of attenuation artifacts in OCT frames reaching F-scores of 0.77 and 0.94 for mild and severe artifacts, respectively. The inference time over a full OCT scan is approximately 6 seconds. Our experiments show that analysis of images represented in both Cartesian and polar coordinate systems outperforms the analysis in polar coordinates only, suggesting that these representations contain complementary features. This work lays the foundation for automated artifact assessment and image acquisition guidance in intracoronary OCT imaging.},

note = {arXiv:2503.05322 [cs]},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{van_der_waerden_artificial_2025,

title = {Artificial intelligence for the analysis of intracoronary optical coherence tomography images: a systematic review},

author = {R. G. A. Waerden and R. H. J. A. Volleberg and T. J. Luttikholt and P. Cancian and J. L. Zande and G. W. Stone and N. R. Holm and E. Kedhi and J. Escaned and D. Pellegrini and G. Guagliumi and S. R. Mehta and N. Pinilla-Echeverri and R. Moreno and L. Räber and T. Roleder and B. Ginneken and C. I. Sánchez and I. Išgum and N. Royen and J. Thannhauser},

doi = {10.1093/ehjdh/ztaf005},

issn = {2634-3916},

year  = {2025},

date = {2025-03-01},

urldate = {2025-03-01},

journal = {European Heart Journal. Digital Health},

volume = {6},

number = {2},

pages = {270–284},

abstract = {Intracoronary optical coherence tomography (OCT) is a valuable tool for, among others, periprocedural guidance of percutaneous coronary revascularization and the assessment of stent failure. However, manual OCT image interpretation is challenging and time-consuming, which limits widespread clinical adoption. Automated analysis of OCT frames using artificial intelligence (AI) offers a potential solution. For example, AI can be employed for automated OCT image interpretation, plaque quantification, and clinical event prediction. Many AI models for these purposes have been proposed in recent years. However, these models have not been systematically evaluated in terms of model characteristics, performances, and bias. We performed a systematic review of AI models developed for OCT analysis to evaluate the trends and performances, including a systematic evaluation of potential sources of bias in model development and evaluation.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{overdevest_model-based_2024,

title = {Model-Based Diffusion for Mitigating Automotive Radar Interference: 49th IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops, ICASSPW 2024},

author = {J. Overdevest and X. Wei and H. Gorp and R. J. G. Sloun},

url = {https://www.scopus.com/pages/publications/85202431169},

doi = {10.1109/ICASSPW62465.2024.10626218},

year  = {2024},

date = {2024-08-01},

urldate = {2024-08-01},

journal = {2024 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops, ICASSPW 2024},

pages = {284–288},

abstract = {Mitigating automotive radar-to-radar interference is a challenging task, especially when the observed signal is densely corrupted with highly correlated interference signals. In this paper, we propose to remove this interference using joint-conditional posterior sampling with score-based diffusion models. These models use three individual scores: a target score, an interference score, and a joint data consistency score. Leveraging the sparsity of clean target signals in the Fourier domain, we propose a model-based score estimator for the target signals, derived from the proximal step of the ℓ1-norm. For the interference score, we use a neural network with denoising score-matching, given that it is difficult to obtain analytical statistical models of the interference signals. Lastly, the target and interference scores are connected by a data-consistency score. Experimental results show that our solution results in superior performance over state-of-the-art methods, in terms of normalized mean squared error (NMSE) and receiver operating characteristic (ROC) curves.},

note = {Publisher: Institute of Electrical and Electronics Engineers},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{li_performance_2024,

title = {Performance Evaluation and Analysis of Thresholding-Based Interference Mitigation for Automotive Radar Systems: 49th IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops, ICASSPW 2024},

author = {J. Li and J. Youn and R. Wu and J. Overdevest and S. Sun},

url = {https://www.scopus.com/pages/publications/85202450787},

doi = {10.1109/ICASSPW62465.2024.10627325},

year  = {2024},

date = {2024-08-01},

urldate = {2024-08-01},

journal = {2024 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops, ICASSPW 2024},

pages = {204–208},

abstract = {In automotive radar, time-domain thresholding (TD-TH) and time-frequency domain thresholding (TFD-TH) are crucial techniques underpinning numerous interference mitigation methods. Despite their importance, comprehensive evaluations of these methods in dense traffic scenarios with different types of interference are limited. In this study, we segment automotive radar interference into three distinct categories. Utilizing the in-house traffic scenario and automotive radar simulator, we evaluate interference mitigation methods across multiple metrics: probability of detection, signal-to-interference-plus-noise ratio, and phase error involving hundreds of targets and dozens of interfering radars. The numerical results highlight that TFD-TH is more effective than TD-TH, particularly as the density and signal correlation of interfering radars escalate.},

note = {Publisher: Institute of Electrical and Electronics Engineers},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{koppelaar_neurally_2024,

title = {Neurally Augmented Deep Unfolding for Automotive Radar Interference Mitigation},

author = {A. G. C. Koppelaar and J. Youn and X. Wei and R. J. G. Sloun},

editor = {J. Overdevest},

doi = {10.1109/TRS.2024.3442692},

issn = {2832-7357},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {IEEE Transactions on Radar Systems},

volume = {2},

number = {10634141},

pages = {712–724},

abstract = {The proliferation of active radar sensors deployed in vehicles has increased the need for mitigating automotive radar-to-radar interference. While simple avoidance and mitigation methods are still effective today, the expected crowded spectrum allocations pose new challenges that likely require more sophisticated techniques. In particular, interference mitigation methods that can handle significant levels of radar signal corruption are required. To this end, we propose neurally augmented analytically learned fast iterative shrinkage thresholding algorithm (NA-ALFISTA), which is a neural network-based solution for reconstructing time-domain radar signals by leveraging sparsity in the range-Doppler map (RDM). The neural augmentation network is deployed as a single gated recurrent unit (GRU) cell that captures the radar signal statistics along the unfolded layers of fast-iterative shrinkage thresholding algorithm (FISTA)-based sparse recovery, which significantly boosts the convergence rate. It estimates the next layer’s parameters necessary in ALFISTA based on the previous layer’s output. The proposed method is compared to state-of-the-art detect-and-repair methods and source separation methods in simulated data and real-world measurements.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{stagnaro_mimo_2023,

title = {MIMO Digital Radar Processing with Spatial Nulling for Self-Interference Mitigation: 2023 IEEE SENSORS, SENSORS 2023},

author = {P. Stagnaro and A. Pandharipande and J. Overdevest and H. Joudeh},

url = {https://www.scopus.com/pages/publications/85179763332},

doi = {10.1109/SENSORS56945.2023.10325195},

year  = {2023},

date = {2023-11-01},

urldate = {2023-11-01},

journal = {2023 IEEE SENSORS},

abstract = {We consider receiver processing in a digital multiple input multiple output (MIMO) radar system in monostatic configuration. The self-interference from transmitter to receiver antenna elements however limits the performance of such systems. We propose analog spatial nulling at the receiver front-end to mitigate the self-interference component. Furthermore, to tackle the Doppler intolerance of binary digital sequences, the proposed receiver processing chain compensates the Doppler shift in fast-time before range processing. We show the improved ability of the proposed system in detecting weak targets.},

note = {Publisher: Institute of Electrical and Electronics Engineers},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{hojjati_training_2025,

title = {Training for Defense: Satellite Events, held together with the 20th International Conference on Persuasive Technology, PERSUASIVE 2025},

author = {A. Hojjati and J. Ham},

editor = {I. Wiafe, A. Babiker, J. Ham, K. Oyibo, E. Vlahu-Gjorgievska},

url = {https://www.scopus.com/pages/publications/105011937417},

doi = {10.1007/978-3-031-97177-8_2},

issn = {978-3-031-97176-1},

year  = {2025},

date = {2025-07-01},

urldate = {2025-07-01},

journal = {Persuasive Technology. PERSUASIVE 2025 Satellite Events},

pages = {18–31},

series = {Communications in Computer and Information Science (CCIS)},

abstract = {Phishing can cause severe security breaches and its frequency and diversity has rapidly increased. Current countermeasures consist mostly of training users in identifying phishing emails by their appearance. However, we argue that in the long run the effect of such trainings will be limited because phishers rapidly evolve their email design and this makes phishing attacks unrecognizable. Still, phishing emails have a universal characteristic: They attempt to influence the users to perform certain behaviors using influencing strategies. Thus, we argue that training users in recognizing the influencing strategies used by technology helps them to defend themselves against (even very advanced, visually unrecognizable) phishing emails. In this study, we randomly assigned 151 participants to two groups (trained on influencing strategies vs. trained on the history of emails). Our learning material was a six-minute training video. After watching the video, participants were presented with a series of emails that contained influencing strategies. These emails were followed by questions about recognition of influencing strategies and the user’s behavioral intentions towards the email. Results provided no evidence that a participant’s intension of clicking on links was influenced by the influencing strategy training video. Importantly, results did show that participants who had watched the influencing strategy training video, correctly recognized more influencing strategies in emails. Also, participants who recognized the use of manipulation techniques in emails, intended to click on less links. These results open a new line of defense against persuasive technology: harnessing users by training them in influencing strategy recognition.},

note = {Place: Cham 

Publisher: Springer},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{ghazali_social_2025,

title = {Social Robots as Decision-Making Companions: Exploring the Impact of Social Cues on Human Responses},

author = {A. S. Ghazali and H. Hafizalshah and S. N. Sidek and H. M. Yusof and J. Ham},

url = {https://www.scopus.com/pages/publications/105005606083},

doi = {10.1142/S0219843625500021},

issn = {0219-8436},

year  = {2025},

date = {2025-06-01},

urldate = {2025-06-01},

journal = {International Journal of Humanoid Robotics},

volume = {22},

number = {3},

abstract = {Making decisions, particularly ones fraught with ambiguity, inherently induces stress, which is a recognized contributor to long-term mental health issues. In high-stakes or uncertain environments, stress can significantly impair decision quality and well-being. Social robots offer a promising solution by potentially providing companionship and cognitive assistance in such scenarios. This study investigates the influence of verbal social cues used by social robots on human responses. In a laboratory setting, 60 participants interacted with the Alpha Mini robot, a programmable social agent, for 30min. The robot offered advice using combinations of controlling language (high versus low) and social praise (absent versus present) in a between-subject design setup while playing a decision-making computer game. Post-interaction, social responses were measured using questionnaires. Results revealed strong, positive correlations between participants’ enjoyment of interacting with the robot and their intention to use it again in the future, as well as their liking and trust in the robot’s advice. These correlations were statistically significant (p<0.01) and suggest that positive user experiences can translate into continued engagement. Positive responses were observed regardless of the specific social cues employed. To design effective human–robot interactions (HRI), multiple social cues should be integrated using high controlling language for clarity and direction paired with social praise to soften the tone in order to enhance trust, enjoyment, and effectiveness. Future work might enhance the current findings by integrating physiological data into the measures used to assess emotional responses to the robot and its cues. Additionally, expanding participant demographics and incorporating longitudinal studies could further validate and extend these results.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{chen_effects_2025,

title = {Effects of Chatbots with Anthropomorphic Visual and Auditory Cues on Users’ Affective Preference: Evidence from Event-Related Potentials},

author = {J. Chen and F. Guo and Z. Zhang and X. Tian and J. Ham},

url = {https://www.scopus.com/pages/publications/105005849554},

doi = {10.1080/10447318.2025.2499169},

issn = {1044-7318},

year  = {2025},

date = {2025-05-01},

urldate = {2025-05-01},

journal = {International Journal of Human-Computer Interaction},

volume = {XX},

abstract = {Anthropomorphic visual and auditory cues are two crucial design elements influencing users’ affective preference for chatbots. However, most earlier studies only focused on one of them and it is still unknown how anthropomorphic visual appearance and voice influence users’ affective preference and neural responses. In the current research, participants’ subjective preference evaluation and objective ERP responses were measured when being presented with chatbots with different visual and auditory cues. (human-like voice and mechanical voice). Subjective results indicated that consistent cues of anthropomorphic visual appearances and voices jointly evoked users’ higher affective preference for chatbots. Notably, auditory cues play a dominant role among audiovisual cues that influence users’ affective preference for chatbots. ERP results showed that low anthropomorphic visual appearances and mechanical voices jointly elicited larger P2 and P3. Additionally, chatbots with low anthropomorphic visual appearances and chatbots with human-like voices elicited larger LPP. These findings hold theoretic implications for understanding the impact of chatbots’ visual appearance and voice on users’ affective preference and provide practical insights for the design of human-chatbot interactions.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{koranteng_impact_2025,

title = {The Impact of Social Support Strategies on Users’ Credibility Perceptions and Continuous Use Intentions of Academic Social Networking Sites: An Empirical Study},

author = {F. N. Koranteng and U. Matzat and I. Wiafe and J. Ham},

url = {https://www.scopus.com/pages/publications/105005531224},

doi = {10.1080/10447318.2025.2495121},

issn = {1044-7318},

year  = {2025},

date = {2025-05-01},

urldate = {2025-05-01},

journal = {International Journal of Human-Computer Interaction},

volume = {XX},

abstract = {With rapid digital innovation, social support strategies are increasingly embedded in Academic Social Networking Sites (ASNSs) to shape user perceptions and behaviors. However, limited empirical research has examined how these strategies influence users’ credibility perceptions and behavioral intentions. Credibility, a key factor driving participation on ASNSs, remains underexplored in this context. This study investigates the role of seven social support strategies within the Persuasive System Design (PSD) framework. It examines their effect on credibility perceptions, as well as the effects of credibility on perceived persuasiveness and continuous use intentions. Using data from 255 ASNS users and Partial Least Squares Structural Equation Modeling (PLS-SEM), results show that several social support strategies significantly shape perceived social learning, which strongly influences credibility perceptions. Additionally, perceived persuasiveness mediates the relationship between perceived credibility and continuous use intention. The findings offer practical insights for designing ASNSs that enhance credibility, persuasiveness, and sustained engagement.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{loman_robustness_2025,

title = {Robustness Measures for Stochastic Parallel Machine Scheduling and Train Unit Shunting: Rail Dresden 2025},

author = {C. Loman and L. Pascual and M. Akker and R. Broek and H. Hoogeveen},

year  = {2025},

date = {2025-03-01},

urldate = {2025-03-01},

journal = {Robustness Measures for Stochastic Parallel Machine Scheduling and Train Unit Shunting},

abstract = {In this paper, we investigate measures that can give us information about the robustness for complex scheduling problems. We identify 14 robustness measures from the literature, as well as introduce 4 new ones. We then use simulation to investigate how well these robustness measures correlate with the stability of the objective function under disturbances (quality robustness), and with the stability of the schedule itself (solution robustness). We first do this in the context of Parallel Machine Scheduling, which is a very general setting that is comparable to many practical situations. We then take the results from that investigation and use the best performing measures as objectives in a local search for the Train Unit Shunting Problem with Service Scheduling. We investigate which of these measures give us a better quality robustness, and which measures give us a better solution robustness. We look at how these measures perform under different ways of inserting slacks into the schedule. We show how the performance of the measures can differ in these different cases, and conclude with what we believe to be a good set of robustness measures to consider for any scheduling problem.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{li_audiovisual_2025,

title = {Audiovisual Affective Design of Humanoid Robot Appearance and Voice Based on Kansei Engineering},

author = {M. Li and J. Zhang and F. Guo and Y. Liao and X. Hu and J. Ham},

url = {https://www.scopus.com/pages/publications/85217195778},

doi = {10.1007/s12369-024-01202-5},

issn = {1875-4791},

year  = {2025},

date = {2025-01-01},

urldate = {2025-01-01},

journal = {International Journal of Social Robotics},

volume = {17},

number = {1},

pages = {15–37},

abstract = {Humanoid robots, characterized by their anthropomorphic design, have become increasingly common in various service areas. Nevertheless, the majority of current affective designs of humanoid robots primarily concentrate on the physical appearance while overlooking its (audiovisual) integration with voice. In this study, we propose simultaneously designing the appearance and voice of humanoid robots using Kansei Engineering, an effective method for optimizing the affective design of products. We first selected representative humanoid robots with different appearances and voices and constructed kansei space to capture users’ affective needs for these robots. Then, we decomposed appearances and parameterized voices to extract design features and orthogonalized these design features to generate prototypes. After that, we conducted an evaluation experiment to acquire users’ affective evaluations on the combinations of appearance and voice. Based on the data, relationship models between design features and users’ kansei images and holistic preferences were constructed using the back-propagation neural network. Furthermore, optimization design models were formulated and resolved through the genetic algorithm. Also, we conducted a validation experiment, and the results demonstrated that the optimized design schemes look harmonious in appearance, sound warmth in voice, and achieve a high level of audiovisual compatibility. The results suggest that the proposed approach can effectively optimize the audiovisual affective design of humanoid robot appearance and voice. Moreover, it can not only provide methodological support for the affective design of robots and other voice-based smart products but can also help to improve the affective experience quality and facilitate the application of robots in service areas.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{koranteng_investigating_2025,

title = {Investigating the Effects of Implicit and Explicit Personalization on Perceived Credibility: 13th International Workshop on Behavior Change Support Systems, BCSS 2025},

author = {F. N. Koranteng and I. Wiafe and J. Ham and U. Matzat},

editor = {H. Oinas-Kukkonen, S. Nabwire},

url = {https://www.scopus.com/pages/publications/105006905426},

year  = {2025},

date = {2025-01-01},

urldate = {2025-01-01},

journal = {Behavior Change Support Systems 2025},

pages = {29–42},

series = {CEUR Workshop Proceedings},

abstract = {Personalizing computer systems (such as Academic Social Networking Sites) can improve positive user perceptions, particularly credibility perceptions of that system. Earlier research has identified two broad personalization approaches: Implicit and Explicit personalization. Moreover, applying the wrong personalization approach may negatively affect users' perceptions of the system's credibility. Yet, the evidence that earlier research provides for the relevance and importance of the different personalization approaches on perceived credibility in system design is limited. This study explores which of the two personalization approaches is most important and could be prioritized when designing systems to improve credibility perceptions. Academic Social Networking Sites (ASNSs) users' perceptions of implicit and explicit personalization and system credibility are gathered via survey and analyzed using Partial Least Square Structural Equation Modeling. We find that whereas Implicit personalization has a positive influence, Explicit personalization negatively influences users' credibility perceptions. Furthermore, the Importance Performance Map Analysis (IPMA) reveals implicit personalization as the better-performing and more important approach for promoting credibility perceptions on ASNSs. Based on the results, this study recommends further investigations into how personalizing the personalization approaches for different users may affect their credibility perceptions.},

note = {Publisher: CEUR-WS.org},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{van_erp_online_2024,

title = {Online Structure Learning with Dirichlet Processes Through Message Passing: 5th International Workshop on Active Inference, IWAI 2024},

author = {B. Erp and W. W. L. Nuijten and B. Vries},

editor = {C. L. Buckley and D. Cialfi and P. Lanillos and R. J. Pitliya and N. Sajid and H. Shimazaki and T. Verbelen and M. Wisse},

url = {https://www.scopus.com/pages/publications/85215817803},

doi = {10.1007/978-3-031-77138-5_6},

issn = {978-3-031-77137-8},

year  = {2024},

date = {2024-12-01},

urldate = {2024-12-01},

journal = {Active Inference},

pages = {91–104},

series = {Communications in Computer and Information Science (CCIS)},

abstract = {Generative or probabilistic modeling is crucial for developing intelligent agents that can reason about their environment. However, designing these models manually for complex tasks is often infeasible. Structure learning addresses this challenge by automating model creation based on sensory observations, balancing accuracy with complexity. Central to structure learning is Bayesian model comparison, which provides a principled framework for evaluating models based on their evidence. This paper focuses on model expansion and introduces an online message passing procedure using Dirichlet processes, a prominent prior in non-parametric Bayesian methods. Our approach builds on previous work by automating Bayesian model comparison using message passing based on variational free energy minimization. We derive novel message passing update rules to emulate Dirichlet processes, offering a flexible and scalable method for online structure learning. Our method generalizes to arbitrary models and treats structure learning identically to state estimation and parameter learning. The experimental results validate the effectiveness of our approach on an infinite mixture model.},

note = {Place: Cham 

Publisher: Springer},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{dameski_system-phenomenology_2024,

title = {System-phenomenology: An empirical case for collectives in mediation theory},

author = {A. Dameski and A. Spahn and C. A. S. Pouw and R. Kodapanakkal and A. Haans and A. Corbetta and F. Toschi and J. R. C. Ham and G. Bombaerts},

doi = {10.59490/jhtr.2024.2.7031},

issn = {2773-2266},

year  = {2024},

date = {2024-12-01},

urldate = {2024-12-01},

journal = {Journal of Human-Technology Relations},

volume = {2},

number = {7031},

abstract = {Postphenomenology and mediation theory strongly explain the micro-level interactions between human individuals and objects. Recently, humans as a collective have been added to the theory at the political macro-level, which we argue that is an important contribution. However, the enlargement of the theory would also merit a meso-level explanation of the role of collectives, in between the micro- and the macro-level. For this purpose, we introduce the mediation triangle, illustrating three bidirectional relations, all mediated by technology: human-object, human-collective, and collective-object. The mediation triangle we combine with three borrowed concepts from systems philosophy to aid in our framework design: differentiality, emergence, and irreducibility. This approach, named system-phenomenology, can explain the interaction between objects, individuals, collectives, political levels, and technology. We illustrate this using an empirical case of boarding and deboarding at train stations. We conclude that system-phenomenology is promising, but further research is needed to develop this theory conceptually.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{chen_different_2024,

title = {Different dimensions of anthropomorphic design cues: How visual appearance and conversational style influence users’ information disclosure tendency towards chatbots},

author = {J. Chen and M. Li and J. Ham},

url = {https://www.sciencedirect.com/science/article/pii/S1071581924001046},

doi = {10.1016/j.ijhcs.2024.103320},

issn = {1071-5819},

year  = {2024},

date = {2024-10-01},

urldate = {2024-10-01},

journal = {International Journal of Human-Computer Studies},

volume = {190},

pages = {103320},

abstract = {Text-based chatbots are widely used to deliver personalized services by leveraging user-provided information, and anthropomorphic design is crucial for their effectiveness. However, most earlier studies investigated the effects of anthropomorphic design of chatbots while manipulating only one dimension of anthropomorphic cues. The current research investigated how different dimensions of anthropomorphic design cues affect users’ information disclosure tendency towards chatbots. That is, the present study examined the effects of visual appearance (high anthropomorphism vs. low anthropomorphism), manipulating the visual cues dimension, and conversational style (human-like vs. mechanical), manipulating the verbal cues dimension, on users’ information disclosure tendency towards chatbots. Results showed positive effects of human-like conversational style on users’ information disclosure tendency. Of particular significance, an interaction effect between visual appearance and conversational style on users’ information disclosure tendency was found. Users reported a higher information disclosure tendency when the chatbot was designed with anthropomorphic cues consistent over dimensions. This finding suggested that an expectancy violation effect occurs when a chatbot exhibits inconsistent anthropomorphic design cues on two different dimensions. Besides, perceived security was identified as a positive mediating factor in the relationship between conversational style and users’ information disclosure tendency. This study advances research on users’ information disclosure tendency towards anthropomorphic chatbots and highlights the importance of different dimensions of anthropomorphic cues in chatbot design. Additionally, practical guidance for chatbot designers was also provided.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{tisza_can_2024,

title = {Can Robots Enhance the Learning Experience by Making Music More Fun?: 33rd IEEE International Conference on Robot and Human Interactive Communication, ROMAN 2024},

author = {G. Tisza and H. Song and P. Markopoulos and E. I. Barakova and J. Ham},

url = {https://www.scopus.com/pages/publications/85209809226},

doi = {10.1109/RO-MAN60168.2024.10731440},

year  = {2024},

date = {2024-10-01},

urldate = {2024-10-01},

journal = {2024 33rd IEEE International Conference on Robot and Human Interactive Communication, ROMAN 2024},

pages = {893–900},

abstract = {Research has shown the potential of social robots to support learning in science, technology, and language. We contribute to this field by exploring how robots can support music learning. We report on a within-subjects experiment where 50 young learners practiced the piano in the presence of a robot assuming a non-evaluative and a self-assessment enhancing role implemented in a Wizard-of-Oz fashion. We examined whether the robot can make piano practice more fun, and whether initiating self-assessment to support self-regulated learning is a useful strategy for the robot. We collected quantitative self-report data to assess fun, learning, interest, engagement, and effort. We found a direct positive effect of fun on learning in the context of musical instrument practice. Path modeling showed a positive influence of having fun on learners' attitudes, interests, and learning outcomes in music education, particularly with the self- assessment robot role exhibiting superiority.},

note = {Publisher: Institute of Electrical and Electronics Engineers},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{kodapanakkal_investigating_2024,

title = {Investigating sociophysical attributes underlying train boarding efficiency and their importance for nudging},

author = {R. I. Kodapanakkal and A. Haans and J. Ham and R. J. Giesen and N. D. Güneş and T. M. L. Markink and J. M. Osinga and C. A. S. Pouw and G. Bombaerts and A. Corbetta and A. Dameski and A. Spahn and F. Toschi},

url = {https://www.scopus.com/pages/publications/85196111014},

doi = {10.1016/j.ssci.2024.106568},

issn = {0925-7535},

year  = {2024},

date = {2024-09-01},

urldate = {2024-09-01},

journal = {Safety Science},

volume = {177},

number = {106568},

abstract = {Nudging has become a popular method to change the behavior of pedestrians in public spaces. However, nudges often do not work as intended because they are based on an incomplete understanding of the nudging environment, physical (e.g., pedestrian trajectories), but not psychological data is used in their development, and behavioral theories are often inadequate or not (correctly) applied. In this article, we argue that the design of nudges can benefit from complementary psychological data analyzed using relevant social and environmental psychological theories. Adequate theories, we argue, are those that aim at describing the objective (i.e., person independent) attributes of the environment or situation and how these affect human decision-making. Using the example of train boarding, and in particular the formation of the deboarding corridor, we demonstrate how psychological theories like interdependence theory and social norms theory can be applied to relevant psychological data—in our case obtained with two focus groups—to better characterize the sociophysical attributes of the train boarding situation. The focus group, or sometimes called a “group discussion”, is a qualitative research method in which data is generated from guided discussions amongst research participants following pre-defined discussion topics. Based on the thematic analysis of the focus group data, we find that a high level of competition and interdependence are related to structural aspects of the train boarding situation. Subsequently, we use these insights to provide tentative explanations for, or hypotheses about micro- and macroscopic behavior patterns observed during train boarding. Finally, we discuss how these insights, in turn, can inform the design of nudges that can be further investigated in future research.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{ruijten_serious_2024,

title = {A serious game for promoting sustainable food choices: 19th International Conference on Persuasive Technology Adjunct, PERSUASIVE-ADJ 2024},

author = {P. A. M. Ruijten and J. Ham and N. Jongh},

editor = {K. Oyibo, W. Xu, E. Vlahu-Gjorgievska},

url = {https://www.scopus.com/pages/publications/85199647853},

year  = {2024},

date = {2024-07-01},

urldate = {2024-07-01},

journal = {Persuasive 2024 Adjunct Proceedings (PERSUASIVE-ADJ 2024)},

pages = {110–111},

series = {CEUR Workshop Proceedings},

abstract = {A type of persuasive technology that has gained popularity over the last decade is gamification. We aimed to influence people’s sustainable food choices by letting them buy ingredients for a dish in two Virtual Reality supermarkets; a gamified one and a regular one. In the gamified supermarket, a point system was added to all the ingredients that could be bought. Also, we used olfactory feedback to enhance the effect of gamification. Results showed an effect of gamification on people’s behavior during the experiment, as well as their self-reported food choices in the week after the experiment. Implications of these findings are discussed in light of how we can persuade people into making healthy and sustainable food choices.},

note = {Publisher: CEUR-WS.org},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{ruijten_how_2024,

title = {How the Role of a Persuasive Robot Impacts One’s Attitude Towards It: 19th International Conference on Persuasive Technology, PERSUASIVE 2024},

author = {P. A. M. Ruijten and J. Smeding and J. Ham},

editor = {N. Baghaei, R. Ali, K. Win, K. Oyibo},

url = {https://www.scopus.com/pages/publications/85192167388},

doi = {10.1007/978-3-031-58226-4_19},

issn = {978-3-031-58225-7},

year  = {2024},

date = {2024-04-01},

urldate = {2024-04-01},

journal = {Persuasive Technology},

pages = {252–261},

series = {Lecture Notes in Computer Science (LNCS)},

abstract = {Recent years have seen a development of social robots in all kinds of different roles. For social robots to be tailored to the needs of the user and become more accepted, we need to understand how people perceive and interact with robots in these different roles. This study investigates people’s attitudes toward robots in two different roles (utilitarian: practically oriented vs hedonic: socially oriented) after interacting with them at home for several days. Results show that people’s attitudes towards the same robot differ between the roles that were applied to the robot. People also described their interactions with the robot in different terms depending on its role. Implications of these findings are discussed in light of tailored approaches in the design of interactions between humans and social robots.},

note = {Place: Cham 

Publisher: Springer},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{chen_effects_2024,

title = {Effects of Anthropomorphic Design Cues of Chatbots on Users’ Perception and Visual Behaviors},

author = {J. Chen and F. Guo and Z. Ren and M. Li and J. Ham},

url = {https://www.scopus.com/pages/publications/85152381313},

doi = {10.1080/10447318.2023.2193514},

issn = {1044-7318},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {International Journal of Human-Computer Interaction},

volume = {40},

number = {14},

pages = {3636–3654},

abstract = {Measurement of users’ perception and visual behaviors to anthropomorphic design cues of chatbots can improve our understanding of chatbots and potentially optimize chatbot design. However, as two typical and basic features, how chatbot appearances and conversational styles jointly affect users’ perception and visual behaviors remains unclear. Therefore, this study conducted an eye-tracking experiment to explore users’ perception and visual behaviors. Results indicate that anthropomorphic appearances and human-like conversational styles jointly increased users’ perception of chatbots’ social presence, trust in chatbots, and satisfaction with chatbots. In contrast, on users’ visual behaviors, such a joint effect was not found, although chatbots with higher anthropomorphic appearances and human-like conversational styles triggered more fixation counts and longer dwell time. These findings suggest that anthropomorphic appearance and human-like conversational style can improve users’ perception and attract more visual attention to chatbots. These findings provide theoretical contributions and practical implications for relevant researchers and designers.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{kodapanakkal_psychological_2024,

title = {A psychological approach to understanding microscopic and macroscopic structures during train boarding processes: International Conference on Traffic and Granular Flow, TGF 2022},

author = {R. I. Kodapanakkal and C. A. S. Pouw and G. Bombaerts and A. Corbetta and A. Dameski and A. Haans and J. Ham and A. Spahn and F. Toschi},

editor = {K. R. Rao and A. Seyfried and A. Schadschneider},

url = {https://www.scopus.com/pages/publications/85197268348},

doi = {10.1007/978-981-99-7976-9_28},

issn = {9789819979752},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {Traffic and Granular Flow'22},

pages = {223–230},

series = {Lecture Notes in Civil Engineering},

abstract = {Current research on the train boarding process focuses predominantly on the physical modelling of pedestrian dynamics and situational characteristics such as platform design. Little psychology is involved in this approach even though individual behavior is a major factor in influencing dwell times. We take a systematic psychological approach to estimate whether parameters like pedestrian speed, area available per pedestrian, and interpersonal distance show variation at the microscopic level (individual variation), macroscopic level (situational variation), or both. Analyzing real-life train (de)boarding events (n = 3728) at a specific location at a Dutch train station, we find that boarders’ speed varies more at the microscopic (individual variation) than the macroscopic level. This behavior could thus result from stable aspects of the situation such as some structural feature of the environment or a social norm. Understanding such variation is helpful in designing behavioral interventions/nudges. If variation is due to individual differences, then an individual-targeted intervention will be most effective. If variation is due to situational differences, then individual-targeted interventions may not be particularly useful, and nudges targeted at crowds or environmental features may be most effective.},

note = {Publisher: Springer},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{song_impact_2024,

title = {The impact of social robots' presence and roles on children's performance in musical instrument practice},

author = {H. Song and E. I. Barakova and J. Ham and P. Markopoulos},

url = {https://onlinelibrary.wiley.com/doi/abs/10.1111/bjet.13416},

doi = {10.1111/bjet.13416},

issn = {1467-8535},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {British Journal of Educational Technology},

volume = {55},

number = {3},

pages = {1041–1059},

abstract = {Research on the educational applications of social robots has shown how they can motivate children and help improve academic learning outcomes. Here, we examine how robots can support skill learning and, more specifically, musical instrument practice. Drawing from social facilitation theory and evaluation apprehension theory we expected that the robot's mere presence would impact children's performance and that this effect would be contingent upon the children expecting the robot to evaluate their performance. We report an experiment with children (N = 31) aged nine to twelve who practiced a familiar and new piece alone, in the presence of an evaluative robot, and in the presence of a non-evaluative robot. We found that children performed better in terms of rhythm, pitch, and general impression in the presence of the non-evaluative robot. These findings offer important insights for designing robot tutors for music learning. Practitioner notes What is already known about this topic Social robots have been applied in different educational scenarios (e.g., second language, math, and programming) and were proven to be beneficial for children's motivation. Musical instrument learning requires practice, perseverance, and social support to become successful. Social robots can be used as a provider of social support during musical instrument practice. Children tend to perform better on easy or well-rehearsed tasks and worse on complex tasks or new ones with the presence of observers, but only when they believe the observer can evaluate them. What this paper adds Social robots are beneficial for children's performance in musical instrument learning. Limited evidence was found to prove that children tend to perform better on old melodies and worse on new melodies in the presence of a social robot. However, the results confirmed that the level of evaluative of the robot matters. Children tend to have better performance with the robot that did not provide evaluative comments when practicing a new melody (a difficult task) than alone and with the robot that offered evaluative comments. This study confirmed that social robots can provide support to children in practicing music, helping to improve their performance. Implications for practice and/or policy Social facilitation and evaluation apprehension effects need to be taken into consideration during the behaviour design of companion robots in learning scenarios. Robots, which were intended to motivate children in learning, should be designed to not provide evaluative comments.},

note = {_eprint: https://bera-journals.onlinelibrary.wiley.com/doi/pdf/10.1111/bjet.13416},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{kodapanakkal_influence_2023,

title = {The Influence of Macroscopic Pedestrian Structures on Train Boarding Efficiency},

author = {R. Kodapanakkal and C. A. S. Pouw and A. Haans and J. R. C. Ham and G. Bombaerts and A. Corbetta and A. Dameski and A. Spahn and F. Toschi},

url = {https://arxiv.org/abs/2309.05476},

doi = {10.48550/arXiv.2309.05476},

year  = {2023},

date = {2023-09-01},

urldate = {2023-09-01},

journal = {The Influence of Macroscopic Pedestrian Structures on Train Boarding Efficiency},

volume = {2309.05476},

pages = {1–27},

abstract = {A deeper understanding of pedestrian dynamics is essential to improve crowd flows in public spaces such as train stations. It is essential to understand both the physical and the psychological processes present in this context. However, current research on train boarding behavior is limited in scope and mainly focuses on how group level variables such as number of boarders/deboarders influence train boarding efficiency. Viewing pedestrian dynamics through a psychological lens is important for a detailed understanding of the train boarding context and to recognize target areas for improving crowd flows. At Dutch train stations, boarders follow a social norm of waiting at the train door until deboarding is complete. Although people generally adhere to this norm, the way it is executed may not be optimal for deboarding efficiency. We investigate how waiting boarders form a deboarding channel (a corridor where deboarders exit the train) which is a macroscopic structure formed by pedestrians, and how this channel in turn influences the efficiency of deboarding. Analyzing a dataset with 3278 boarding events at Utrecht Centraal Station in the Netherlands from 2017 - 2020 (a subset of a trajectory dataset that captures 100,000 trajectories per day), we found that higher numbers of boarders and a higher ratio of boarders to deboarders, reduced the width of the deboarding channel, and a lower width was associated with lower deboarding efficiency. These results shift the focus from group level variables to identifying macroscopic structures that are formed when pedestrians interact within a social system and provide specific target areas where nudges/behavioral interventions could be implemented.},

note = {Publisher: arXiv.org},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{oyibo_adaptive_2023,

title = {Adaptive and Personalized Persuasive Technologies (ADAPPT 2023) Workshop: 31st ACM Conference on User Modeling, Adaptation and Personalization, UMAP 2023},

author = {K. Oyibo and I. Adaji and R. Orji and J. Ham and J. Vassileva},

url = {https://www.scopus.com/pages/publications/85163702688},

doi = {10.1145/3563359.3595626},

year  = {2023},

date = {2023-06-01},

urldate = {2023-06-01},

journal = {UMAP '23 Adjunct},

pages = {121–122},

abstract = {The Adaptive and Personalized Persuasive Technologies (ADAPPT'23) workshop holding in Cyprus this year is the third edition of the ADAPPT series, which commenced in 2019. The workshop is organized in conjunction with the 31st Association for Computer Machinery (ACM) Conference on User Modeling, Adaptation and Personalization (UMAP). In this preface to the third edition, we summarize the papers accepted for publication in the adjunct proceedings. Finally, we present a list of the members of the organizing and program committees that made the workshop a success.},

note = {Publisher: Association for Computing Machinery, Inc.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{song_how_2023,

title = {‘How Would you Score Yourself?’: The Effect of Self-assessment Strategy Through Robots on Children’s Motivation and Performance in Piano Practice},

author = {H. Song and K. Tsiakas and J. Ham and P. Markopoulos and E. I. Brakova},

url = {http://www.scopus.com/inward/record.url?scp=85180207523&partnerID=8YFLogxK},

doi = {10.1007/s12369-023-01080-3},

issn = {1875-4791},

year  = {2023},

date = {2023-01-01},

urldate = {2023-01-01},

journal = {International Journal of Social Robotics},

volume = {16},

number = {2},

pages = {327–340},

abstract = {This research examines how to design social robots to support self-regulated learning skills for piano practice. More specifically, a social robot is used to provide feedback to children and initiate self-assessment. To assess the impact of this approach on children’s motivation and performance, we conducted an experiment in a music school where 50 children practiced with both a self-assessment and a non-evaluative robot. Results showed that when the children interacted with the self-assessment robot they had higher motivation and better performance than when they interacted with the non-evaluative robot. Furthermore, interaction effects were found between the robot conditions, the children’s learning stages, and their gender regarding their motivation and rhythm performance. Overall, the study demonstrates a positive influence of robot-initiated self-assessment on children’s musical instrument practice and provided insights for personalized child-robot interaction design.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{tocchetti_i_2025,

title = {Aİ. Robustness: a Human-Centered Perspective on Technological Challenges and Opportunities},

author = {A. Tocchetti and L. Corti and A. Balayn and M. Yurrita and P. Lippmann and M. Brambilla and J. Yang},

url = {https://dl.acm.org/doi/10.1145/3665926},

doi = {10.1145/3665926},

issn = {0360-0300},

year  = {2025},

date = {2025-02-01},

urldate = {2025-02-01},

journal = {ACM Comput. Surv.},

volume = {57},

number = {6},

pages = {141:1–141:38},

abstract = {Despite the impressive performance of Artificial Intelligence (AI) systems, their robustness remains elusive and constitutes a key issue that impedes large-scale adoption. Besides, robustness is interpreted differently across domains and contexts of AI. In this work, we systematically survey recent progress to provide a reconciled terminology of concepts around AI robustness. We introduce three taxonomies to organize and describe the literature both from a fundamental and applied point of view: (1) methods and approaches that address robustness in different phases of the machine learning pipeline; (2) methods improving robustness in specific model architectures, tasks, and systems; and in addition, (3) methodologies and insights around evaluating the robustness of AI systems, particularly the tradeoffs with other trustworthiness properties. Finally, we identify and discuss research gaps and opportunities and give an outlook on the field. We highlight the central role of humans in evaluating and enhancing AI robustness, considering the necessary knowledge they can provide, and discuss the need for better understanding practices and developing supportive tools in the future.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{arzberger_nothing_2024,

title = {Nothing Comes Without Its World – Practical Challenges of Aligning LLMs to Situated Human Values through RLHF},

author = {A. Arzberger and S. Buijsman and M. L. Lupetti and A. Bozzon and J. Yang},

url = {https://ojs.aaai.org/index.php/AIES/article/view/31617},

doi = {10.1609/aies.v7i1.31617},

issn = {3065-8365},

year  = {2024},

date = {2024-10-01},

urldate = {2024-10-01},

journal = {Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society},

volume = {7},

number = {1},

pages = {61–73},

abstract = {Work on value alignment aims to ensure that human values 

are respected by AI systems. However, existing approaches 

tend to rely on universal framings of human values that obscure 

the question of which values the systems should capture 

and align with, given the variety of operational situations. 

This often results in AI systems that privilege only a 

selected few while perpetuating problematic norms grounded 

on biases, ultimately causing equity and justice issues. In this 

perspective paper, we unpack the limitations of predominant 

alignment practices of reinforcement learning from human 

feedback (RLHF) for LLMs through the lens of situated values. 

We build on feminist epistemology to argue that at the 

design-time, RLHF has problems with representation in the 

subjects providing feedback and implicitness in the conceptualization 

of values and situations of real-world users while 

lacking system adaptation to real user situations at the use time. 

To address these shortcomings, we propose three research 

directions: 1) situated annotation to capture information 

about the crowdworker’s and user’s values and judgments 

in relation to specific situations at both the design and 

use-time, 2) expressive instruction to encode plural values 

for instructing LLMs systems at design-time, and 3) reflexive 

adaptation to leverage situational knowledge for system 

adaption at use-time. We conclude by reflecting on the 

practical challenges of pursuing these research directions and 

situated value alignment of AI more broadly.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{biswas_hi_nodate,

title = {“Hi. I’m Molly, Your Virtual Interviewer!” — Exploring the Impact of Race and Gender in AI-powered Virtual Interview Experiences},

author = {S. Biswas and J. Jung and A. Unnam and K. Yadav and S. Gupta and U. Gadiraju},

url = {https://arxiv.org/html/2408.14159v1},

year  = {2024},

date = {2024-08-26},

urldate = {2024-08-26},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{sun_revisiting_2024,

title = {Revisiting Bundle Recommendation for Intent-aware Product Bundling},

author = {Z. Sun and K. Feng and J. Yang and H. Fang and X. Qu and Y. S. Ong and W. Liu},

url = {https://dl.acm.org/doi/10.1145/3652865},

doi = {10.1145/3652865},

year  = {2024},

date = {2024-06-01},

urldate = {2024-06-01},

journal = {ACM Trans. Recomm. Syst.},

volume = {2},

number = {3},

pages = {24:1–24:34},

abstract = {Product bundling represents a prevalent marketing strategy in both offline stores and e-commerce systems. Despite its widespread use, previous studies on bundle recommendation face two significant limitations. Firstly, they rely on noisy datasets, where bundles are defined by heuristics, e.g., products co-purchased in the same session. Secondly, they target specific tasks by holding unrealistic assumptions, e.g., the availability of bundles for recommendation directly. This paper proposes to take a step back and considers the process of bundle recommendation from a holistic user experience perspective. We first construct high-quality bundle datasets with rich metadata, particularly bundle intents, through a carefully designed crowd-sourcing task. We then define a series of tasks that together, support all key steps in a typical bundle recommendation process, from bundle detection, completion and ranking, to explanation and auto-naming, whereby 19 research questions are raised correspondingly to guide the analysis. Finally, we conduct extensive experiments and analyses with representative recommendation models and large language models (LLMs), demonstrating the challenges and opportunities, especially with the emergence of LLMs. To summarize, our study contributes by introducing novel data sources, paving the way for new research avenues, and offering insights to guide product bundling in real e-commerce platforms.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{yang_fedtrans_2024,

title = {FedTrans: Client-Transparent Utility Estimation for Robust Federated Learning},

author = {M. Yang and R. Zhu and Q. Wang and J. Yang},

url = {https://proceedings.iclr.cc/paper_files/paper/2024/hash/bb309cc1fbdb88ea755bd7cee4b310ec-Abstract-Conference.html},

year  = {2024},

date = {2024-05-01},

urldate = {2024-05-01},

journal = {International Conference on Representation Learning},

volume = {2024},

pages = {42668–42692},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{smirnova_xcrowd_2024,

title = {XCrowd: 33rd ACM International Conference on Information and Knowledge Management, CIKM 2024},

author = {A. Smirnova and J. Yang and P. Cudre-Mauroux},

url = {http://www.scopus.com/inward/record.url?scp=85210019001&partnerID=8YFLogxK},

doi = {10.1145/3627673.3679777},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {CIKM '24},

pages = {2097–2107},

abstract = {Relation extraction methods are currently dominated by deep neural models, which capture complex statistical patterns while being brittle and vulnerable to perturbations in data and distribution. Explainability techniques offer a means for understanding such vulnerabilities, and thus represent an opportunity to mitigate future errors; yet, existing methods are limited to describing what the model 'knows', while totally failing at explaining what the model does not know. This paper presents a new method for diagnosing model predictions and detecting potential inaccuracies. Our approach involves breaking down the problem into two components: (i) determining the necessary knowledge the model should possess for accurate prediction, through human annotations, and (ii) assessing the actual knowledge possessed by the model, using explainable AI methods (XAI). We apply our method to several relation extraction tasks and conduct an empirical study leveraging human specifications of what a model should know and does not know. Results show that human workers are capable of accurately specifying the model should-knows, despite variations in the specification, that the alignment between what a model really knows and what it should know is indeed indicative of model accuracy, and that the unknowns identified through our methods allow to foresee future errors that may or may not have been observed otherwise.},

note = {Place: New York, NY 

Publisher: ACM},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{salimzadeh_when_2024,

title = {When in Doubt! Understanding the Role of Task Characteristics on Peer Decision-Making with AI Assistance: 32nd ACM Conference on User Modeling, Adaptation and Personalization},

author = {S. Salimzadeh and U. Gadiraju},

url = {http://www.scopus.com/inward/record.url?scp=85197883809&partnerID=8YFLogxK},

doi = {10.1145/3627043.3659567},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {UMAP 2024 - Proceedings of the 32nd ACM Conference on User Modeling, Adaptation and Personalization},

pages = {89–101},

series = {UMAP 2024 - Proceedings of the 32nd ACM Conference on User Modeling, Adaptation and Personalization},

abstract = {With the integration of AI systems into our daily lives, human-AI collaboration has become increasingly prevalent. Prior work in this realm has primarily explored the effectiveness and performance of individual human and AI systems in collaborative tasks. While much of decision-making occurs within human peers and groups in the real world, there is a limited understanding of how they collaborate with AI systems. One of the key predictors of human-AI collaboration is the characteristics of the task at hand. Understanding the influence of task characteristics on human-AI collaboration is crucial for enhancing team performance and developing effective strategies for collaboration. Addressing a research and empirical gap, we seek to explore how the features of a task impact decision-making within human-AI group settings. In a 2 × 2 between-subjects study (N = 256) we examine the effects of task complexity and uncertainty on group performance and behaviour. The participants were grouped into pairs and assigned to one of four experimental conditions characterized by varying degrees of complexity and uncertainty. We found that high task complexity and high task uncertainty can negatively impact the performance of human-AI groups, leading to decreased group accuracy and increased disagreement with the AI system. We found that higher task complexity led to a higher efficiency in decision-making, while a higher task uncertainty had a negative impact on efficiency. Our findings highlight the importance of considering task characteristics when designing human-AI collaborative systems, as well as the future design of empirical studies exploring human-AI collaboration.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{he_opening_2024,

title = {Opening the Analogical Portal to Explainability: Can Analogies Help Laypeople in AI-assisted Decision Making?},

author = {G. He and A. Balayn and S. Buijsman and J. Yang and U. Gadiraju},

url = {http://www.scopus.com/inward/record.url?scp=85204874277&partnerID=8YFLogxK},

doi = {10.1613/jair.1.15118},

issn = {1076-9757},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {Journal of Artificial Intelligence Research},

volume = {81},

pages = {117–162},

abstract = {Concepts are an important construct in semantics, based on which humans understand the world with various levels of abstraction. With the recent advances in explainable artificial intelligence (XAI), concept-level explanations are receiving an increasing amount of attention from the broad research community. However, laypeople may find such explanations difficult to digest due to the potential knowledge gap and the concomitant cognitive load. Inspired by prior work that has explored analogies and sensemaking, we argue that augmenting concept-level explanations with analogical inference information from commonsense knowledge can be a potential solution to tackle this issue. To investigate the validity of our proposition, we first designed an effective analogy-based explanation generation method and collected 600 analogy-based explanations from 100 crowd workers. Next, we proposed a set of structured dimensions for the qualitative assessment of such explanations, and conducted an empirical evaluation of the generated analogies with experts. Our findings revealed significant positive correlations between the qualitative dimensions of analogies and the perceived helpfulness of analogy-based explanations, suggesting the effectiveness of the dimensions. To understand the practical utility and the effectiveness of analogybased explanations in assisting human decision-making, we conducted a follow-up empirical study (N = 280) on a skin cancer detection task with non-expert humans and an imperfect AI system. Thus, we designed a between-subjects study spanning five different experimental conditions with varying types of explanations. The results of our study confirmed that a knowledge gap can prevent participants from understanding concept-level explanations. Consequently, when only the target domain of our designed analogy-based explanation was provided (in a specific experimental condition), participants demonstrated relatively more appropriate reliance on the AI system. In contrast to our expectations, we found that analogies were not effective in fostering appropriate reliance. We carried out a qualitative analysis of the open-ended responses from participants in the study regarding their perceived usefulness of explanations and analogies. Our findings suggest that human intuition and the perceived plausibility of analogies may have played a role in affecting user reliance on the AI system. We also found that the understanding of commonsense explanations varied with the varying experience of the recipient user, which points out the need for further work on personalization when leveraging commonsense explanations. In summary, although we did not find quantitative support for our hypotheses around the benefits of using analogies, we found considerable qualitative evidence suggesting the potential of high-quality analogies in aiding non-expert users in their decision making with AI-assistance. These insights can inform the design of future methods for the generation and use of effective analogy-based explanations.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{corti_it_2024,

title = {“It Is a Moving Process”: 2024 CHI Conference on Human Factors in Computing Sytems, CHI 2024},

author = {L. Corti and R. Oltmans and J. Jung and A. Balayn and M. Wijsenbeek and J. Yang},

editor = {F. Mueller and P. Kyburz and J. R. Williamson and C. Sas and M. L. Wilson and P. Toups Dugas and I. Shklovski},

url = {http://www.scopus.com/inward/record.url?scp=85194835253&partnerID=8YFLogxK},

doi = {10.1145/3613904.3642551},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

journal = {CHI '24: Proceedings of the CHI Conference on Human Factors in Computing Systems},

abstract = {Clinicians increasingly pay attention to Artificial Intelligence (AI) to improve the quality and timeliness of their services. There are converging opinions on the need for Explainable AI (XAI) in healthcare. However, prior work considers explanations as stationary entities with no account for the temporal dynamics of patient care. In this work, we involve 16 Idiopathic Pulmonary Fibrosis (IPF) clinicians from a European university medical centre and investigate their evolving uses and purposes for explainability throughout patient care. By applying a patient journey map for IPF, we elucidate clinicians' informational needs, how human agency and patient-specific conditions can influence the interaction with XAI systems, and the content, delivery, and relevance of explanations over time. We discuss implications for integrating XAI in clinical contexts and more broadly how explainability is defined and evaluated. Furthermore, we reflect on the role of medical education in addressing epistemic challenges related to AI literacy.},

note = {Publisher: ACM},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{khaled_interpretable_2025,

title = {Interpretable Fuzzy Systems For Forward Osmosis Desalination: 2025 IEEE International Conference on Fuzzy Systems, FUZZ IEEE 2025},

author = {Q. Khaled and U. Kaymak and L. Genga},

url = {https://www.scopus.com/pages/publications/105017427277},

doi = {10.1109/FUZZ62266.2025.11152221},

year  = {2025},

date = {2025-09-01},

urldate = {2025-09-01},

journal = {2025 IEEE International Conference on Fuzzy Systems, FUZZ 2025},

abstract = {Preserving interpretability in fuzzy rule-based systems (FRBS) is vital for water treatment, where decisions impact public health. While structural interpretability has been addressed using multi-objective algorithms, semantic interpretability often suffers due to fuzzy sets with low distinguishability. We propose a human-in-the-loop approach for developing interpretable FRBS to predict forward osmosis desalination productivity. Our method integrates expert-driven grid partitioning for distinguishable membership functions, domain-guided feature engineering to reduce redundancy, and rule pruning based on firing strength. This approach achieved comparable predictive performance to cluster-based FRBS while maintaining semantic interpretability and meeting structural complexity constraints, providing an explainable solution for water treatment applications.},

note = {Publisher: Institute of Electrical and Electronics Engineers},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{latifi_federated_2025,

title = {A Federated Game-Theoretic Strategy for Three-Phase Voltage Unbalance Mitigation Via Dynamic Pricing and Demand Response textbar Request PDF},

author = {M. Latifi and C. Zucca and M. C. A. Willemsen},

url = {https://www.researchgate.net/publication/394173947_A_Federated_Game-Theoretic_Strategy_for_Three-Phase_Voltage_Unbalance_Mitigation_Via_Dynamic_Pricing_and_Demand_Response},

doi = {10.2139/ssrn.5374774},

year  = {2025},

date = {2025-08-01},

urldate = {2025-08-01},

journal = {ResearchGate},

abstract = {Request PDF textbar On Jan 1, 2025, Milad Latifi and others published A Federated Game-Theoretic Strategy for Three-Phase Voltage Unbalance Mitigation Via Dynamic Pricing and Demand Response textbar Find, read and cite all the research you need on ResearchGate},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{khaled_optimizing_2025,

title = {Optimizing Takagi-Sugeno Fuzzy Models For Improving Leak Detection in Water Distribution Networks: IEEE Symposium Series on Computational Intelligence IEEE-SSCI 2025},

author = {Q. Khaled and U. Kaymak and L. Genga},

url = {https://www.scopus.com/pages/publications/105007717050},

doi = {10.1109/CIETES63869.2025.10995166},

year  = {2025},

date = {2025-05-01},

urldate = {2025-05-01},

journal = {2025 IEEE Symposia on Computational Intelligence for Energy, Transport and Environmental Sustainability, CIETES},

abstract = {Leakage detection in water distribution networks (WDNs) is critical for reducing water loss and ensuring operational efficiency. While machine learning methods are often applied, they can lack interpretability. Takagi-Sugeno (Tsk) fuzzy systems offer a balance between accuracy and interpretability but are prone to overfitting and incur high computational costs, especially with large datasets. To address these issues, we explore various optimization and regularization techniques to improve Tsk performance. The models were trained on a large-scale benchmark dataset containing 1000 leak scenarios, each a year-long time series at half-hour intervals, totaling over 17 million data points. A systematic preprocessing pipeline was applied, including time-series segmentation, mutual information-based feature selection, and class imbalance handling. Alongside the baseline training of Adaptive Neuro-Fuzzy Inference Systems (ANFIS) using gradient descent (GD), we also trained Tsk models using stochastic gradient descent (SGD) and mini-batch gradient descent (MBGD). State-of-the-art regularization techniques such as uniform regularization and rule dropout were also incorporated to prevent overfitting. Key results show that SGD and MBGD models outperformed GD models in leak detection rates and achieved significantly lower false alarm rates than traditional machine learning models. These findings underscore the potential of fuzzy systems for effective leak detection, provided that appropriate learning and regularization techniques are employed.},

note = {Publisher: Institute of Electrical and Electronics Engineers},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{khaled_end–end_2025,

title = {An End-to-End Framework for AI Integration in Desalination Systems: 20th International Conference on Information Processing and Management of Uncertainty in Knowledge-Based Systems, IPMU2024},

author = {Q. Khaled and U. Kaymak and L. Genga},

editor = {M. J. Lesot, S. Vieira, M. Reformat, F. Batista, J. P. Carvalho, B. Bouchon-Menier, R. R. Yager},

doi = {10.5281/zenodo.15149423},

year  = {2025},

date = {2025-04-01},

urldate = {2025-04-01},

journal = {IPMU2024 Lisboa - Short Paper Proceedings},

pages = {39–43},

abstract = {In the context of population growth, urbanization, and industrial expansion, water scarcity emerges as a significant concern, with projections indicating that around two billion individuals may face this challenge by 2050. Hence, the increased pressure on existing water resources calls for new water supply solutions in light of the growing demand. Desalination emerges as a promising alternative solution, particularly in regions confronting limited water resources. The sector has experienced remarkable growth, witnessing a 41% capacity increase over the past decade, with projections hinting at a twofold expansion by 2030. Such expansion requires integrating cutting-edge modeling techniques to ensure efficacy and cost-effectiveness. Artificial intelligence (AI) shows potential to revolutionize desalination and water treatment practices, yet its implementation remains limited. Delayed integration is believed to stem from the lack of trust among domain experts, knowledge gaps between water professionals and data scientists, and untapped potential within the field. This paper proposes The Integrated System Perspective for AI-based Desalination (ISP); an End-to-End Framework for AI in desalination. ISP-AID facilitates identifying AI applications across various project stages, from design to maintenance, uncovering opportunities for cost reduction and efficiency improvement. It adopts a structured data science perspective, integrating the Cross-Industry Standard Process for Data Mining (CRISP-DM) to guide AI algorithm selection and deployment. Spanning project cycle, process design, and data science levels, the framework aims to instill trust, foster collaborative problem understanding, and highlight untapped potential. This positions domain experts to actively develop data-driven solutions and enhancing confidence in innovative methodologies. By facilitating collaboration and exploring AI applications, the framework could expedite adopting efficient desalination solutions, thereby addressing global water scarcity challenges.},

note = {Publisher: Zenodo},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{khaled_semi-supervised_2025,

title = {A semi-supervised graph-based approach for anomaly detection in river monitoring stations: NCR Days 2025 - Crossing Boundaries},

author = {Q. Khaled and U. Kaymak and L. Genga},

editor = {V. Chavarrias, A. M. Hoek},

year  = {2025},

date = {2025-01-01},

urldate = {2025-01-01},

journal = {Crossing boundaries},

volume = {57-2025},

pages = {28–29},

abstract = {Water quality monitoring is crucial for ensuring safe drinking water and protecting aquatic ecosystems. Traditional methods, like periodic sampling, often fail to detect sudden pollution events, delaying responses that could protect public health and the environment Zhu et al. (2022). Anomalies, such as nutrient spikes or oxygen drops, can signal environmental issues like pollution or runoff, necessitating immediate detection. This gap highlights the need for advanced techniques capable of realtime data analysis. Machine learning, particularly graph neural networks (GNNs), has shown promise in enhancing real-time water quality monitoring. GNNs are effective at modeling the spatial connections between monitoring stations, which helps in predicting water quality more accurately Li et al. (2024) Yan and Wang (2024). This is especially useful for understanding how pollution spreads across a network. However, a significant hurdle is the scarcity of labeled data for training GNNs, as anomalies in water quality are rare and often not well-documented Buchhorn et al. (2024). This makes supervised learning challenging, pushing researchers toward novel methods to detect unusual events without extensive labeled datasets.},

note = {Publisher: Netherlands Centre for River Studies},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{obreja_characterizing_2025,

title = {Characterizing the Impact of Training Data on Generalizability: Application in Deep Learning to Estimate Lung Nodule Malignancy Risk},

author = {B. Obreja and J. Bosma and K. V. Venkadesh and Z. Saghir and M. Prokop and C. Jacobs},

url = {https://pubs.rsna.org/doi/10.1148/ryai.240636},

doi = {10.1148/ryai.240636},

year  = {2025},

date = {2025-11-01},

urldate = {2025-11-01},

journal = {Radiology: Artificial Intelligence},

volume = {7},

number = {6},

pages = {e240636},

abstract = {Purpose To investigate the relationship between training data volume and performance of a deep learning artificial intelligence (AI) algorithm developed to assess the malignancy risk of pulmonary nodules detected on low-dose CT scans in lung cancer screening.Materials and Methods This retrospective study used a dataset of 16 077 annotated nodules (1249 malignant, 14 828 benign) from the National Lung Screening Trial (NLST) to systematically train an AI algorithm for pulmonary nodule malignancy risk prediction across various stratified subsets ranging from 1.25% to the full dataset. External testing was conducted using data from the Danish Lung Cancer Screening Trial (DLCST) to determine the amount of training data at which the performance of the AI was statistically noninferior to the AI trained on the full NLST cohort. A size-matched cancer-enriched subset of DLCST, in which each malignant nodule had been paired in diameter with the closest two benign nodules, was used to investigate the amount of training data at which the performance of the AI algorithm was statistically noninferior to the average performance of 11 clinicians.Results The external testing set included 599 participants (mean age ± SD, 57.65 years ± 4.84 for female participants and 59.03 years ± 4.94 for male participants) with 883 nodules (65 malignant, 818 benign). The AI achieved a mean area under the receiver operating characteristic curve (AUC) of 0.92 (95% CI: 0.88, 0.96) on the DLCST cohort when trained on the full NLST dataset. Training with 80% of the NLST data resulted in noninferior performance (mean AUC, 0.92; 95% CI: 0.89, 0.96; P = .005). On the size-matched DLCST subset (59 malignant, 118 benign), the AI reached noninferior clinician-level performance (mean AUC, 0.82; 95% CI: 0.77, 0.86) with 20% of the training data (P = .02).Conclusion The deep learning AI algorithm demonstrated excellent performance in assessing pulmonary nodule malignancy risk, achieving clinical level performance with a fraction of the training data and reaching peak performance before using the full dataset. Keywords: Convolutional Neural Network (CNN), CT, Lung, Screening, Diagnosis, Supervised Learning, Lung Cancer Screening, Pulmonary Nodule Malignancy Risk, Deep Learning, Pulmonary Nodule Management Supplemental material is available for this article. © RSNA, 2025 See also commentary by Archer in this issue.},

note = {Publisher: Radiological Society of North America},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{vitale_beyond_nodate,

title = {Beyond ‘artificial intelligence’: against anthropomorphizing algorithmic systems for screening},

author = {M. Vitale},

url = {https://www.tijdschrifttge.nl/art/50-8606_Beyond-artificial-intelligence-against-anthropomorphizing-algorithmic-systems-for-screening},

year  = {2025},

date = {2025-09-16},

urldate = {2025-09-16},

abstract = {As a PhD candidate in the ethics of AI for lung cancer screening at MERAI lab, I have the opportunity to conduct my research within an interdisciplinary context. Working alongside engineers and...},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{hanou_multi-agent_2025,

title = {Multi-Agent Pathfinding for Railway Routing: RailDresden 2025: 11th International Conference on Railway Operations Modelling and Analysis},

author = {I. Hanou and M. de Weerdt},

url = {https://tu-dresden.de/bu/verkehr/die-fakultaet/veranstaltungen/raildresden2025/ressourcen/dateien/BoA_RailDresden_full_final.pdf?lang=en},

year  = {2025},

date = {2025-01-01},

urldate = {2025-01-01},

pages = {102–102},

abstract = {Research in railway operations has mostly focused on operations research methods. However, these real-world problems have a state-based nature, which makes them very suitable for AI models, such as the Multi-Agent Pathfinding problem, where agents move in a grid and need to be routed from their start to their goal location without colliding with each other. The core aspect of problems like train shunting and train dispatching is routing, which is often not the main focus of current mathematical formulations. Therefore, we apply the state-of-the-art algorithms to the railway problems of shunting and dispatching and study their usability for routing trains. The Multi-Agent Pathfinding problem is often solved with one of two algorithms: conflict-based search (a two-stage algorithm detecting conflicts between individual paths and using A* search to find new conflict-free paths), and branch-cut-and-price (a linear program adding cuts (row generation) based on problem-specific constraints, and finding new paths to be selected that satisfy all constraints using a pricer). We modify these algorithms to include more railway details. First, we allow for the matching of train units (i.e., ensure the necessary train units of a certain type are available for departure) by specifying goals for agent (type) groups instead of single agent goals. Moreover, we add goal sequences for servicing stations and agents of different sizes, and we study specific aspects of the railway infrastructure to exploit in the algorithm. Finally, we show the use of Multi-Agent Pathfinding solvers in different railway settings and analyze the conditions for success.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{kemmeren_introducing_2025,

title = {Introducing flexibility in any-start-time safe interval path planning},

author = {E. Kemmeren},

url = {https://repository.tudelft.nl/record/uuid:04890fe5-d983-4b91-90bd-dc5da7129161},

year  = {2025},

date = {2025-01-01},

urldate = {2025-01-01},

abstract = {During the daily operation of the railway network, ProRail is responsible for handling delays and planning ad hoc train movements. Train handling documents aid the traffic controllers in common situations. But when multiple trains are delayed, and these documents do not apply, they are left to their own expertise. <br/>In this thesis, we introduce FlexSIPP, an algorithm to plan or replan agents in an existing multi-agent plan. FlexSIPP builds upon the prior works of any-start-time safe interval path planning, where the current routes of the agents are seen as moving obstacles. FlexSIPP loosens this restriction by introducing flexibility: the ability for an agent to delay its plan while minimally impacting other agents. <br/>This algorithm is evaluated on the Dutch railway network. By finding tipping points, that is, the moment it is better to switch the order of two trains on the track to minimize the delay, we can recreate train handling documents. We show that FlexSIPP finds the same solutions within a minute in the case that no other trains are delayed. This implies that FlexSIPP is also able to aid traffic controllers in the case that other trains are delayed},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{hanou_replanning_2024,

title = {Replanning in Advance for Instant Delay Recovery in Multi-Agent Applications: Rerouting Trains in a Railway Hub},

author = {I. K. Hanou and D. W. Thomas and W. Ruml and M. de Weerdt},

url = {https://ojs.aaai.org/index.php/ICAPS/article/view/31483},

doi = {10.1609/icaps.v34i1.31483},

issn = {2334-0843},

year  = {2024},

date = {2024-05-01},

urldate = {2024-05-01},

journal = {Proceedings of the International Conference on Automated Planning and Scheduling},

volume = {34},

pages = {258–266},

abstract = {Train routing is sensitive to delays that occur in the network. When a train is delayed, it is imperative that a new plan be found quickly, or else other trains may need to be stopped to ensure safety, potentially causing cascading delays. In this paper, we consider this class of multi-agent planning problems, which we call Multi-Agent Execution Delay Replanning. We show that these can be solved by reducing the problem to an any-start-time safe interval planning problem. When an agent has an any-start-time plan, it can react to a delay by simply looking up the precomputed plan for the delayed start time. We identify crucial real-world problem characteristics like the agent's speed, size, and safety envelope, and extend the any-start-time planning to account for them. Experimental results on real-world train networks show that any-start-time plans are compact and can be computed in reasonable time while enabling agents to instantly recover a safe plan.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{lonyuk_using_2024,

title = {Using PDDL models to solve TUSS},

author = {N. Lonyuk},

url = {https://repository.tudelft.nl/record/uuid:242e8fdd-95d2-4915-bd28-f0697559514e},

year  = {2024},

date = {2024-01-01},

urldate = {2024-01-01},

abstract = {Due to the increased demand for train travel, train operators are considering increasing their rolling stock. Before achieving this, they must enhance the capacity of their shunting yards. This is attempted by improving methodologies for solving the Train Unit Shunting and Servicing (TUSS) problem. To address the TUSS problem, a planner determines routes on shunting yards for trains, ensuring they visit designated service tracks before parking in a configuration that facilitates a smooth departure. <br/>TUSS is a well-studied problem, and various approaches have been proposed. The first approach capable of solving real-world, complete TUSS instances is a local search method introduced by van den Broek et al. In this thesis, we explore an alternative approach using PDDL models. PDDL is the standard language for describing Automated Planning problems. Automated Planning is a well-established field within artificial intelligence, and new, improved algorithms are continually developed to solve PDDL models for problems similar to TUSS.<br/>In this thesis, we design a detailed model in PDDL and propose several methods to simplify the model so that planning algorithms perform more efficiently compared to the detailed model. When solving simplified models, a post-processing routine is employed to generate detailed shunting plans. The performance of several model-independent PDDL planners was analysed, and the best-performing planner was identified as Temporal FastDownward. <br/>By analysing plans obtained from experiments, we identified areas for improvement. Based on this knowledge, we developed a new TUSS-specific planner called Train Order Preserving Search (TOPS). TOPS employs a search algorithm with effective pruning of symmetrical states and a custom heuristic that guides the search towards states where the order of trains aligns with the departure order. TOPS significantly outperformed Temporal FastDownward in these experiments.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{bocsardi_moderating_2025,

title = {The Moderating Effects of Atypicality on Content Performance in the Online Mass Media Industry},

author = {Á. G. Bocsárdi and K. Podoynitsyna and C. Zucca},

url = {https://journals.aom.org/doi/10.5465/AMPROC.2025.15138abstract},

doi = {10.5465/AMPROC.2025.15138abstract},

issn = {0065-0668},

year  = {2025},

date = {2025-07-01},

urldate = {2025-07-01},

journal = {Academy of Management Proceedings},

volume = {2025},

number = {1},

pages = {15138},

abstract = {Media holding companies face a fundamental tension between leveraging economies of scale through content sharing and maintaining distinctive brand identities across their portfolios. While content sharing across outlets offers clear operational benefits, it risks diluting unique brand voices that attract loyal audiences. This study examines how atypicality — standing out among similar entities, both in terms of content characteristics and network position — affects this trade-off. Drawing on 32 months of data from a European media conglomerate encompassing 1.5M articles shared across a network of 17 news and magazine brands, our results confirm that increased content adoption volume generally reduces content performance. Atypicality of the content and the way brands adopt articles from each other (i.e., content-level and network-level atypicality) can turn this into an advantage. We find that relying on more atypical content adoption strategies can help media organizations offset the negative effects of increased adoption volume. These findings advance our understanding of how atypicality functions in consumption- and experience-based digital markets and offer practical insights for media organizations balancing efficiency with differentiation.},

note = {Publisher: Academy of Management},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{li_hypertension_2025,

title = {Hypertension, intracranial arteriosclerosis, and structural brain changes in patients with TIA or ischemic stroke},

author = {X. Li and B. P. Berghout and G. Rooijen and M. K. Ikram and B. Roozenbeek and D. Bos},

doi = {10.1177/23969873241307099},

issn = {2396-9881},

year  = {2025},

date = {2025-09-01},

urldate = {2025-09-01},

journal = {European Stroke Journal},

volume = {10},

number = {3},

pages = {804–812},

abstract = {INTRODUCTION: Hypertension is a major risk factor of structural brain changes, including atrophy and cerebral small vessel disease. Intracranial arteriosclerosis could be an underlying mechanism between hypertension and structural brain changes. This study investigated whether intracranial carotid artery calcification (ICAC), as a proxy for intracranial arteriosclerosis, explains the association between hypertension and structural brain changes in patients with TIA or ischemic stroke. 

PATIENTS AND METHODS: About 968 patients (mean age 62.7 years) with TIA or ischemic stroke from a registry who underwent non-contrast CT (NCCT) and CT-angiography (CTA) were included in this study. Presence and volume (mm3) of ICAC were assessed on CTA. Subtypes of ICAC were assessed on NCCT, where ICAC was categorized into intimal and internal elastic lamina (IEL) type calcification. Structural brain changes, indicated by atrophy, periventricular and deep white matter lesions (WML), and lacunes were assessed on NCCT. Mediation analysis was performed using ICAC, ICAC volume, and ICAC subtypes as mediators. 

RESULTS: ICAC was prevalent in 67.8% of patients, with 52.6% of them exhibiting intimal calcification, and 26.5% exhibiting IEL calcification. Atrophy, periventricular WML, deep WML, and lacunes were present in 48.1%, 56.4%, 43.0% and 17.1% of patients respectively. The presence of ICAC explained 7.1% of the association of hypertension with periventricular WML, 3.6% with deep WML, and 17.6% with lacunes. Hypertension was associated with increased atrophy through ICAC (OR: 1.02, 95% CI: 1.00-1.05). In subgroup analyses, IEL calcification partly explained the association between hypertension and periventricular WML (16.8%), and atrophy (OR: 1.12, 95% CI: 1.02-1.27). Intimal calcification did not explain any association. 

CONCLUSION: ICAC partially explained the association between hypertension and atrophy, periventricular and deep WML, and lacunes. Although intimal calcification was more prevalent in ischemic stroke patients, IEL calcification takes the leading role in explaining the association between hypertension and structural brain changes.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{van_hulst_decision-analytic_2025,

title = {A Decision-Analytic Model to Evaluate Cost-Effectiveness of Regional Implementation of a Mobile Stroke Unit},

author = {P. L. Hulst and R. M. Wijdeven and E. Venema and F. M. E. Pinckaers and M. G. M. Hunink and A. Lugt and D. W. J. Dippel and H. F. Lingsma and D. Bos and B. Roozenbeek},

doi = {10.1212/WNL.0000000000213834},

issn = {1526-632X},

year  = {2025},

date = {2025-08-01},

urldate = {2025-08-01},

journal = {Neurology},

volume = {105},

number = {3},

pages = {e213834},

abstract = {BACKGROUND AND OBJECTIVES: Mobile stroke units (MSUs) have the potential to improve functional outcome of ischemic stroke patients, through shortening onset-to-treatment times. Previous cost-effectiveness studies have limited generalizability to nonmetropolitan settings and did not evaluate cost-effectiveness over a lifetime horizon. We aimed to develop a regionally adaptable decision-analytic model, to evaluate cost-effectiveness of MSU implementation and to identify the optimal dispatch scenario. 

METHODS: We developed a generalizable state-transition microsimulation model with modifiable region-specific parameters and dispatch characteristics to evaluate the lifetime cost-effectiveness from a health care perspective of 1-year MSU implementation. We used the southwest of the Netherlands (1,770,000 inhabitants, 1,592 km2, 7 primary stroke centers, 2 thrombectomy-capable stroke centers) as an example. Region-specific input parameters for the model, such as population density, age distribution, and driving times, were obtained at the level of postal codes. We developed a virtual cohort of suspected stroke patients based on age-dependent stroke risks and the number of inhabitants per postal code. We compared the combined dispatch of an MSU and emergency medical services (EMS) with dispatch of EMS alone for patients with onset-to-alarm time <6 hours, living within the catchment area of the MSU. In the base case analysis, the MSU could be dispatched to all postal codes in the study region between 7.00 am and 11.00 pm from a central dispatch site. We assessed the long-term cost-effectiveness through incremental net monetary benefits (iNMBs). Discount rates were 1.5% for effects and 4.0% for costs. 

RESULTS: In the base case scenario, the MSU was dispatched to 2,080 of 3,628 patients (57.3%) with a suspected stroke and onset-to-alarm time <6 hours, resulting in a lifetime gain of 399 (95% CI 384-414) additional quality-adjusted life years, €3.9 million (95% CI €3.5 million-€4.3 million) cost savings, and an iNMB of €23.9 million (95% CI €22.8 million-€24.9 million). A smaller catchment area for MSU dispatch was associated with increased cost-effectiveness. 

DISCUSSION: Adding an MSU to the dispatch strategy for suspected stroke patients is expected to be cost-effective in our region. Our model facilitates evaluation of the cost-effectiveness of MSU implementation in different regions, settings, and scenarios with varying characteristics.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{kang_rethinking_2025,

title = {Rethinking Click Models in Light of Carousel Interfaces: Theory-Based Categorization and Design of Click Models},

author = {J. Kang and M. de Rijke and S. Leon-Martinez and H. Oosterhuis},

url = {https://dl.acm.org/doi/10.1145/3731120.3744585},

doi = {10.1145/3731120.3744585},

year  = {2025},

date = {2025-07-01},

urldate = {2025-07-01},

journal = {Proceedings of the 2025 International ACM SIGIR Conference on Innovative Concepts and Theories in Information Retrieval (ICTIR)},

pages = {44–55},

series = {ICTIR '25},

abstract = {Click models are a well-established for modeling user interactions with web interfaces. Previous work has mainly focused on traditional single-list web search settings; this includes existing surveys that introduced categorizations based on the first generation of probabilistic graphical model (PGM) click models that have become standard. However, these categorizations have become outdated, as their conceptualizations are unable to meaningfully compare PGM with neural network (NN) click models nor generalize to newer interfaces, such as carousel interfaces. We argue that this outdated view fails to adequately explain the fundamentals of click model designs, thus hindering the development of novel click models. This work reconsiders what should be the fundamental concepts in click model design, grounding them - unlike previous approaches - in their mathematical properties. We propose three fundamental key-design choices that explain what statistical patterns a click model can capture, and thus indirectly, what user behaviors they can capture. Based on these choices, we create a novel click model taxonomy that allows a meaningful comparison of all existing click models; this is the first taxonomy of single-list, grid and carousel click models that includes PGMs and NNs. Finally, we show how our conceptualization provides a foundation for future click model design by an example derivation of a novel design for carousel interfaces.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{de_leon-martinez_recgaze_2025,

title = {RecGaze: The First Eye Tracking and User Interaction Dataset for Carousel Interfaces},

author = {S. Leon-Martinez and J. Kang and R. Moro and M. de Rijke and B. Kveton and H. Oosterhuis and M. Bielikova},

url = {https://dl.acm.org/doi/10.1145/3726302.3730301},

doi = {10.1145/3726302.3730301},

year  = {2025},

date = {2025-07-01},

urldate = {2025-07-01},

journal = {Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval},

pages = {3702–3711},

series = {SIGIR '25},

abstract = {Carousel interfaces are widely used in e-commerce and streaming services, but little research has been devoted to them. Previous studies of interfaces for presenting search and recommendation results have focused on single ranked lists, but it appears their results cannot be extrapolated to carousels due to the added complexity. Eye tracking is a highly informative approach to understanding how users click, yet there are no eye tracking studies concerning carousels. There are very few interaction datasets on recommenders with carousel interfaces and none that contain gaze data. We introduce the RecGaze dataset: the first comprehensive feedback dataset on carousels that includes eye tracking results, clicks, cursor movements, and selection explanations. The dataset comprises of interactions from 3 movie selection tasks with 40 different carousel interfaces per user. In total, 87 users and 3,477 interactions are logged. In addition to the dataset, its description and possible use cases, we provide results of a survey on carousel design and the first analysis of gaze data on carousels, which reveals a golden triangle or F-pattern browsing behavior. Our work seeks to advance the field of carousel interfaces by providing the first dataset with eye tracking results on carousels. In this manner, we provide and encourage an empirical understanding of interactions with carousel interfaces, for building better recommender systems through gaze information, and also encourage the development of gaze-based recommenders.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{nachesa_knn_2025,

title = {kNN For Whisper And Its Effect On Bias And Speaker Adaptation},

author = {M. K. Nachesa and V. Niculae},

editor = {L. Chiruzzo and A. Ritter and L. Wang},

url = {https://aclanthology.org/2025.findings-naacl.369/},

doi = {10.18653/v1/2025.findings-naacl.369},

year  = {2025},

date = {2025-04-01},

urldate = {2025-04-01},

journal = {Findings of the Association for Computational Linguistics: NAACL 2025},

pages = {6621–6627},

abstract = {Speech recognition performance varies by language, domain, and speaker characteristics such as accent, but fine-tuning a model on any of these categories may lead to catastrophic forgetting. Token-level k nearest neighbor search (kNN), first proposed for neural sequence decoders for natural language generation (NLG) and machine translation (MT), is a non-parametric method that instead adapts using inference-time search in an external datastore, without training the underlying model. We show that Whisper, a transformer end-to-end speech model, benefits from kNN. We investigate the differences between the speech and text setups. We discuss implications for speaker adaptation, and analyze improvements by gender, accent, and age.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{ganesh_does_2024,

title = {Does SpatioTemporal information benefit Two video summarization benchmarks?},

author = {A. Ganesh and M. Popa and D. Odijk and N. Tintarev},

url = {http://arxiv.org/abs/2410.03323},

doi = {10.48550/arXiv.2410.03323},

year  = {2024},

date = {2024-10-01},

urldate = {2024-10-01},

abstract = {An important aspect of summarizing videos is understanding the temporal context behind each part of the video to grasp what is and is not important. Video summarization models have in recent years modeled spatio-temporal relationships to represent this information. These models achieved state-of-the-art correlation scores on important benchmark datasets. However, what has not been reviewed is whether spatio-temporal relationships are even required to achieve state-of-the-art results. Previous work in activity recognition has found biases, by prioritizing static cues such as scenes or objects, over motion information. In this paper we inquire if similar spurious relationships might influence the task of video summarization. To do so, we analyse the role that temporal information plays on existing benchmark datasets. We first estimate a baseline with temporally invariant models to see how well such models rank on benchmark datasets (TVSum and SumMe). We then disrupt the temporal order of the videos to investigate the impact it has on existing state-of-the-art models. One of our findings is that the temporally invariant models achieve competitive correlation scores that are close to the human baselines on the TVSum dataset. We also demonstrate that existing models are not affected by temporal perturbations. Furthermore, with certain disruption strategies that shuffle fixed time segments, we can actually improve their correlation scores. With these results, we find that spatio-temporal relationship play a minor role and we raise the question whether these benchmarks adequately model the task of video summarization. Code available at: https://github.com/AashGan/TemporalPerturbSum},

note = {arXiv:2410.03323 [cs]},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{zilbershtein_bridging_2024,

title = {Bridging the Transparency Gap: Exploring Multi-Stakeholder Preferences for Targeted Advertisement Explanations},

author = {D. Zilbershtein and F. Barile and D. Odijk and N. Tintarev},

url = {http://arxiv.org/abs/2409.15998},

doi = {10.48550/arXiv.2409.15998},

year  = {2024},

date = {2024-09-01},

urldate = {2024-09-01},

abstract = {Limited transparency in targeted advertising on online content delivery platforms can breed mistrust for both viewers (of the content and ads) and advertisers. This user study (n=864) explores how explanations for targeted ads can bridge this gap, fostering transparency for two of the key stakeholders. We explore participants' preferences for explanations and allow them to tailor the content and format. Acting as viewers or advertisers, participants chose which details about viewing habits and user data to include in explanations. Participants expressed concerns not only about the inclusion of personal data in explanations but also about the use of it in ad placing. Surprisingly, we found no significant differences in the features selected by the two groups to be included in the explanations. Furthermore, both groups showed overall high satisfaction, while "advertisers" perceived the explanations as significantly more transparent than "viewers". Additionally, we observed significant variations in the use of personal data and the features presented in explanations between the two phases of the experiment. This study also provided insights into participants' preferences for how explanations are presented and their assumptions regarding advertising practices and data usage. This research broadens our understanding of transparent advertising practices by highlighting the unique dynamics between viewers and advertisers on online platforms, and suggesting that viewers' priorities should be considered in the process of ad placement and creation of explanations.},

note = {arXiv:2409.15998 [cs]},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{richterich_feminist_2024,

title = {Feminist automation: Can bots have feminist politics?},

author = {A. Richterich and S. Wyatt},

url = {https://www.scopus.com/pages/publications/85202795927},

doi = {10.1177/14614448241251801},

issn = {1461-4448},

year  = {2024},

date = {2024-09-01},

urldate = {2024-09-01},

journal = {New Media & Society},

volume = {26},

number = {9},

pages = {4973–4991},

abstract = {This article examines ‘feminist chatbots’ as tools for activism through automation. Such bots aim to engage users in automated communication on feminist concerns. The article starts from the assumption that chatbots, like all technologies, have politics and that automation, including the automated communication of chatbots, is a feminist issue. We investigate how feminist chatbots mobilise automation to address societal inequalities and bias. Conceptually, the article draws on technofeminism and intersectionality as lenses for understanding the potential of chatbots to reflect activist concerns. Three different chatbots are analysed, using a cultural (case) studies approach: F’xa, Gender Pay Gap Bot and Betânia. The analysis suggests that feminist chatbots oppose mainstream automation by engaging users in communication about its sociotechnical risks and using automation to inspire feminist (data) activism. Yet challenges remain in designing such bots, partly because of platform dependencies and the limits of automating complex intersectional issues.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{kang_estimating_2024,

title = {Estimating the Hessian Matrix of Ranking Objectives for Stochastic Learning to Rank with Gradient Boosted Trees},

author = {J. Kang and M. de Rijke and H. Oosterhuis},

url = {https://dl.acm.org/doi/10.1145/3626772.3657918},

doi = {10.1145/3626772.3657918},

year  = {2024},

date = {2024-07-01},

urldate = {2024-07-01},

journal = {Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval},

pages = {2390–2394},

series = {SIGIR '24},

abstract = {Stochastic learning to rank (LTR) is a recent branch in the LTR field that concerns the optimization of probabilistic ranking models. Their probabilistic behavior enables certain ranking qualities that are impossible with deterministic models. For example, they can increase the diversity of displayed documents, increase fairness of exposure over documents, and better balance exploitation and exploration through randomization. A core difficulty in LTR is gradient estimation, for this reason, existing stochastic LTR methods have been limited to differentiable ranking models (e.g., neural networks). This is in stark contrast with the general field of LTR where Gradient Boosted Decision Trees (GBDTs) have long been considered the state-of-the-art. In this work, we address this gap by introducing the first stochastic LTR method for GBDTs. Our main contribution is a novel estimator for the second-order derivatives, i.e., the Hessian matrix, which is a requirement for effective GBDTs. To efficiently compute both the first and second-order derivatives simultaneously, we incorporate our estimator into the existing PL-Rank framework, which was originally designed for first-order derivatives only. Our experimental results indicate that stochastic LTR without the Hessian has extremely poor performance, whilst the performance is competitive with the current state-of-the-art with our estimated Hessian. Thus, through the contribution of our novel Hessian estimation method, we have successfully introduced GBDTs to stochastic LTR.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}

@article{tintarev_measuring_2024,

title = {Measuring the benefit of increased transparency and control in news recommendation},

author = {N. Tintarev and B. P. Knijnenburg and M. C. Willemsen},

url = {https://www.scopus.com/pages/publications/85190986468},

doi = {10.1002/aaai.12171},

issn = {0738-4602},

year  = {2024},

date = {2024-06-01},

urldate = {2024-06-01},

journal = {AI Magazine},

volume = {45},

number = {2},

pages = {212–226},

abstract = {Personalized news experiences powered by recommender systems permeate our lives and have the potential to influence not only our opinions, but also our decisions. At the same time, the content and viewpoints contained within news recommendations are driven by multiple factors, including both personalization and editorial selection. Explanations could help users gain a better understanding of the factors contributing to the news items selected for them to read. Indeed, recent works show that explanations are essential for users of news recommenders to understand their consumption preferences and set intentions in line with their goals, such as goals for knowledge development and increased diversity of content or viewpoints. We give examples of such works on explanation and interactive interface interventions which have been effective in influencing readers' consumption intentions and behaviors in news recommendations. However, the state-of-the-art in news recommender systems currently fall short in terms of evaluating such interventions in live systems, limiting our ability to measure their true impact on user behavior and opinions. To help understand the true benefit of these interfaces, we therefore call for improving the realism of studies for news.},

keywords = {},

pubstate = {published},

tppubtype = {article}

}