ICAI interview with Georgios Vlassopoulos: Strengthening the friendship between AI and humans

Georgios Vlassopoulos is a PhD student at KPN Responsible AI Lab, located in Den Bosch. He works on explainability systems for AI-models. Vlassopoulos: ‘If we want artificial intelligence predicting stuff, we should be able to explain why it makes a certain decision. Without this transparency, AI could be dangerous.’

Georgios Vlassopoulos

What is your research about?

‘My algorithm tries to explain to the user why the AI-system has made a certain decision. I’ll give an example. KPN uses natural language processing on texts looking for complaints of customers. How do you explain to the user why the system has categorized certain texts? What is the decision of the computer based on? In this case, my algorithm tries to learn the semantics that people use in complaints to use in the explanation of the model.’

‘You can expand this to many domains. Say that a doctor uses AI to detect cancer. The doctor only sees the prediction of the model, so if the patient has cancer or not. The patient must be very eager to know why the computer has made a certain decision. With my algorithm, I would learn the attributes, like the shape of a tumour, to the system and build an understandable explanation based on these attributes.’

How do you approach your research?

‘Let’s stick to the KPN example. For a large amount of texts the classifier would say: I’m not sure if it’s a complaint or not. I focus on the decision boundary, which is the set of datapoints for which the classifier is completely uncertain. All the classification information is encoded in this decision boundary, which is very complex. My approach is to train a simple model which mimics the behaviour of only a certain part of this complex information. And this can be communicated to the user.’

Why is your research different from other methods?

‘The explanations of current popular explaining models can be misleading. When you use these methods on high dimensional data, e.g. images, they treat every pixel as an individual feature. My position is that you cannot build a proper explanation based on pixels. I introduced a different framework that scales well for high dimensional data. And the explanations become more humanlike.’

Why is your research important?

‘In a data-driven world it is very important for AI to become friends with human beings. People should be able to understand why an AI-system makes a certain decision. If a bank classifies its customers with an AI-system on whether they are fit to receive a loan or not, then they should be able to inform the customers why they are accepted or rejected.’

What are the main challenges you face doing this research?

‘It’s like you’re looking for aliens. There is no ground truth. The problem is that you don’t really have an accuracy measure. If we take the medical example, a doctor can say that an explanation from the system is close to his intuition. But how can you prove that this is actually correct? I need to design the experiments carefully and still everything can go wrong. Sometimes I have to repeat an experiment multiple times.’

What are you most proud of?

‘The fact that I have made something that works. And it has good chances to be published in a top conference. The final answer will come in January 2021. But I’m already proud that high impact scientists say that my work is good.’

In this months’ Lunch at ICAI Meetup on Transparency and Trust in AI on December 17 Georgios Vlassopoulos will discuss his research. Sign up here.

3 factors for a successful university-business collaboration

Mark Siebert, Director of Research Collaborations at Elsevier, wrote an interesting piece on Elsevier Connect about the ingredients that make a collaboration between academia and industry work. He used his personal impressions and lessons learned from establishing Elsevier’s Discovery Lab. 3 success factors played a role he says:

  1. Build credibility, reliability and shared excitement as you get to know each other’s unique abilities.
  2. Differentiate business/research priorities and legal requirements to minimize “self-orientation.”
  3. Use facilitated communication to connect the interests and expectations that must converge to build trust.

‘There is nothing better in a partnership than winning together’, Siebert concludes.

Read the whole article here

Prepare to Integrate AI in Your Organization – New Online Program

How can your organization benefit from AI? The new online program AI in Practice, starting on November 3, helps professionals to implement AI in their own organization. This program has been developed in collaboration with the Labs of ICAI and is offered on the edX.org platform.

Just like electricity must have been more than 100 years ago, AI is now a disruptive force with the potential to revolutionize industries and businesses across the board. We may not yet be able to properly imagine how it will affect our work and lives, but organizations need to start thinking about how AI can be applied to improve everyday practice.

The program consists of 2 short online courses aimed at anyone who, regardless of their professional background or job role, is interested in learning how AI can help their organizations solve specific problems, drive efficiency and improve performance and decision-making.

Is this then a program for you or for someone you know? Visit our program page and help us spread the word by sharing this with your networks, colleagues and anyone interested in learning how AI can help their organizations and how to develop a plan for its application.

Start date: November 3, 2020

Snoek op Zolder – New podcast by NL AIC and ICAI

Now listen to the second episode of ‘Snoek op Zolder’, starring professor Frank van Harmelen. Snoek op Zolder is the new weekly science podcast by the Dutch AI Coalition (NL AIC) and ICAI.

In this second episode, Hosts Hennie Huijgens and Sietse van Gorkum speak with Frank van Harmelen (VU), professor knowledge representation and reasoning, on the Discovery Lab. Within this lab, Elsevier, VU and UvA, conduct research into the use of AI to support scientist in their research methods. They discuss the twenty million euros, ten-year collaboration between researchers from six Dutch universities in the Hybrid Intelligence Center and the research lag of Europe compared to America and China.

Special treat: the song ‘Artificial Intelligence’ written exclusively for this podcast by singer-songwriter Yanna Pelser.

About de podcast

In Snoek op Zolder the hosts will be discussing small and large AI innovations, exciting research results and earth-shaking theories every week with passionate curiosity and humor. With this podcast they aim to spread knowledge about AI and inspire people to learn more about it.

There will be a new podcast every week, you can listen to the first episode on any podcast app or platform!

Launching of the sixteenth ICAI-lab: AI for Agro-Food lab

Together with the 4TU, OnePlanet and several industry partners, Wageningen University & Research has set up the AI for Agro-Food lab. This sixteenth ICAI lab will evolve a new generation of ecology-based sustainable agricultural production systems that are supported and enhanced by smart tailored technology. The lab is the first ICAI-lab in Wageningen.

The AI for Agro-Food lab will provide methods and tools that form the base for both integrated solutions for production systems, as well as sociotechnical scenarios for the transition towards these systems. This lab will be a subset of the larger Synergia crossover program initiated by NWO.

Research topics

One of the research topics of the lab will be smart crop and animal sensing. The researchers aim to develop dedicated sensors for applications in horticulture, dairy farming and arable farming. Secondly, the lab will focus on intelligent control and decision making by developing sustainable methods for customizable autonomous decision-making and management support of specific farming challenges. Finally, robotic handling will be an important research topic. The researchers will focus on developing algorithms to enable the accomplishment of the robot actuation tasks.

The lab will be under the academic supervision of professor Peter Groot Koerkamp. Yannick van Gelder is lab manager.

Read more on AI for Agro-Food lab.

City of Amsterdam, VU and UvA launch a lab for social artificial intelligence

The Civic AI Lab is a collaboration between the City of Amsterdam, Vrije Universiteit Amsterdam (VU) and the University of Amsterdam (UvA). The College has approved a subsidy of €400,000 for the coming four years, while the VU will contribute €900,000 and the UvA €1.5 million. Within the lab, five researchers will work towards obtaining their PhD while studying the application of artificial intelligence in the fields of education, welfare, the environment, mobility and health. The lab will form part of the Innovation Center for Artificial Intelligence (ICAI) at Amsterdam Science Park.

Artificial intelligence (AI) is the collective term for applications in which technology makes use of algorithms to perform tasks that would typically require human intelligence. AI is playing an increasingly important role in our society. For example, businesses use AI to improve their production processes, to select job applicants and to analyse health data. Throughout the coronavirus outbreak, AI has helped to visualise the progress of the pandemic – Amsterdam UMC, for example, uses an algorithm to map the spread of the coronavirus in the Netherlands.

However, there are also disadvantages to deploying AI. Image recognition systems, for example, were typically trained using images of people with a lighter skin tone and therefore were not as accurate when recognising those with darker skin tones. Insurers also use mathematical models that identify postcodes where there is an increased risk of burglary. However, this results in higher premiums for the residents of these neighbourhoods. As a result, this increases inequality in a city.

With the Civic AI Lab, the City wants to examine examples of such friction so that in the future AI will promote equality and deliver fair opportunities, overcoming its negative side effects. Deputy Mayor Touria Meliani (Digital City) says: “It is the City of Amsterdam’s ambition to be a leader in the application of AI and its ethical implementation. Technology and AI have many benefits, but they can also put basic rights such as privacy, freedom of speech and legal equality under pressure. The City wants to protect these rights to the best of its ability and the Civic AI Lab should contribute to this.”

Deputy Mayor Victor Everhardt (Economic Affairs) says: “Work in the lab will include developing algorithms that responsibly bring together data flows about births and youth healthcare so that we can derive predictive factors from them. Models are also being developed to ensure a good and fair distribution of finances for pupils and schools. I view artificial intelligence as the science of the future, but it does carry certain risks. That is why it is so important that ethical issues are investigated in the lab.”

In the Civic AI Lab, five PhD students will conduct research on AI issues in the areas of education, welfare, the environment, mobility and health. Scientific director Sennay Ghebreab (UvA) says: “Working within these five topics, the lab will develop AI technology that highlights the inequality of opportunity in society and that actively increases the prospect of equality of opportunity.” The lab will also serve as an information point for residents and businesses who have questions about new technologies and the ethical and inclusive use of them. The lab should make residents aware of both the opportunities and risks of AI.

Find out more about Civic AI Lab.

TU/e-EAISI and industry define new ICAI-lab for “AI-enabled manufacturing and maintenance”

Together with a number of industrial partners, Eindhoven University of Technology (TU/e) has set up a new ICAI-lab to improve decision making in manufacturing and maintenance using artificial intelligence. The new EAISI lab “AI-enabled Manufacturing and Maintenance” (AIMM) is the first ICAI-Lab in Eindhoven. 

TU/e has had a multi-disciplinary Smart Manufacturing and Maintenance research program for several years,
in order to develop better data science techniques and information from complex systems to improve
manufacturing and maintenance. With the establishment of TU/e’s AI institute EAISI last year and the
growing trend of using AI to accomplish even more sophisticated tasks, it is an obvious evolution of the
program to establish the EAISI AI-enabled Manufacturing and Maintenance (AIMM) lab.

AIMM aims to promote research in cooperation with high-tech industry and is a collaboration with the four industrial
partners Nexperia, KMWE, Marel and Lely. These companies have a solid history of commitment to developing new innovative approaches in their respective industries and are looking forward to working in close partnership with the TU/e.

Important step

Carlo van de Weijer, director of EAISI, welcomes the establishment of the new ICAI lab as an important step.
“We are proud that we can contribute towards a responsible and ethical use of AI. This cooperation will have
benefits for innovation, development of talent and will result in a much stronger role for AI in the Dutch

As part of the initiative, seven PhD research positions will perform joint research at the business partners
and the TU/e. The areas of research will directly interface with topics such as Autonomous Agents &
Robotics, Computer Vision, Decision Making, information Retrieval, Knowledge Representation & Reasoning,
NLP and Machine Learning. AIMM Lab will be under the academic supervision of Ivo Adan and Geert-Jan van Houtum.

It is expected that the EAISI AIMM research lab will also become a fertile starting point for new initiatives,
such as developing online trainings and placing Master and final year Bachelor students in projects with
companies. The lab will also provide companies with better access to expertise in potential applications of
artificial intelligence such as improved planning, detecting anomalies in production, predicting quality issues
and enabling fast root cause analysis.

Read more on AIMM Lab

Opening Donders AI for Neurotech Lab

Radboud University is investing further in Artificial Intelligence with the opening of a third ICAI lab at Radboud University. It is the thirteenth ICAI lab in the Netherlands. The Donders AI for Neurotech Lab is working on the development of next-generation of AI-driven neurotechnologies. Its aim is to improve the life of patients who rely on neurotech solutions to overcome or ease their disabilities.

The AI for Neurotech Lab aims to develop machine learning solutions for brain reading and writing, to restore sensory and cognitive functions. These solutions could include hearing tools for the deaf and vision for the blind, but also tools that oppress seizures for those that suffer from epilepsy. Yağmur Güçlütürk, manager of the lab and assistant professor Artificial Intelligence at Radboud University, explains: ‘An example of such work, which I am very proud of, is our ongoing national and international collaborations on developing the first long-term visual cortical implants to restore visual perception in blind people.’


The lab is a partnership between the Donders Institute for Brain, Cognition and Behavior, several companies (including Advanced Bionics, Abbott, Phosphoenix and the OnePlanet Research Center) and various academic institutions from across the Netherlands. To support new research at the lab, five PhD-students will be hired shortly to help develop new AI solutions, made possible by 14 million euros of funding through the NWO INTENSE CrossOver program.

Marcel van Gerven, scientific director of the lab: ‘The Donders ICAI AI for Neurotech lab will develop new AI solutions that power next-generation neurotechnology. Our goal is to restore communication and control in people who lost this ability due to a wide variety of impairments that affect brain function.’

Güçlütürk: ‘We hope to make the Donders AI for Neurotech Lab a central hub in Europe for developing smarter neurotech. To do this, we will be building the lab with several partners from both the academic world and the industry.’

Radboud AI

The lab is part of Radboud AI, an initiative that brings together researchers from all seven faculties of Radboud University. In total, the university has over 170 scientists working on research that involves artificial intelligence. Based on the citation score, it is the number one in AI research in the Netherlands. The university distinguishes itself with a focus on the human side of artificial intelligence, with a lot of attention on the applications of AI where humans are at the center.

Take a look at the website of Donders AI for Neurotech Lab for more information.

JADS & KPN launch responsible AI Lab: Using AI in a Responsible Way

‘s-Hertogenbosch, May 15, 2020 – ICAI is joining forces with leading Dutch telecommunications and IT provider KPN and Jheronimus Academy of Data Science (JADS), for the launch of a new public-private research lab. The KPN Responsible AI Lab will focus on developing technologies that allow Artificial Intelligence (AI) to be used in a responsible way.

The lab will be based in ‘s-Hertogenbosch and will be one of the [ICAI](https://icai.ai) labs; the national network on technology and talent development between universities, industries and government in the area of Artificial Intelligence.

“We are extremely proud to welcome JADS and KPN with the KPN Responsible AI Lab as new partners in our national ICAI network. It is a very significant step that JADS and KPN join forces to help advance the development and use of AI in the Netherlands,” said Maarten de Rijke, Scientific Director of ICAI.

KPN and JADS work together on PhD research trajectories in the areas where data science, engineering, decision-making, privacy and compliance are all combined. The results of which will be used in practice to realize responsible AI solutions for the benefit of the company and its clients and the larger society.

JADS is the joint answer of the Technical University of Eindhoven and Tilburg University to the strongly increasing demand for experts who can improve decision-making processes within organizations, based on big data and a multi-disciplinary approach. Eric Postma, Scientific Director at the KPN-JADS Responsible AI lab comments: *”Data science is all about technology (mathematics, statistics and computer science), but always within the context of domain expertise on the one hand and social, ethical and legal issues on the other.”*

The goal of the Responsible AI lab, is to develop transparent, privacy aware, and personalized AI solutions for businesses.

For the next five years, six PhD students will work in the KPN Responsible AI lab on projects contributing to the technical and legal aspects of transparency, on data engineering approaches that ensure privacy-by-design, on personalized and responsible offerings through text analytics and responsible up-lift modelling.

“Within KPN we are committed in applying AI in a responsible way,” said Winifred Andriessen , Director Advanced Analytics at KPN. “Therefore we need multidisciplinary research both in AI technology as well as in how to use this technology in a societal optimal way. This is why we’re collaborating with JADS; a multidisciplinary institute combining technology and social sciences. This will move us a step closer to a company that not only has a deep understanding of the potential power of AI, but also a company that is able to transform this power into added value for its customers and the broader society like energy efficiency and fibre rollout.”

More information about Responsible AI Lab of JADS & KPN

AIM Lab starts with a grand opening

The official opening of the AI for Medical Imaging lab (AIM Lab) on Friday 31 January, was performed by Geert ten Dam (President of the Executive Board of the University of Amsterdam), Victor Everhardt (Alderman of the City of Amsterdam) and Xiantong Zhen (IIAI) by revealing the banner of the AIM Lab logo.

AIM Lab is a research collaboration between the Inception Institute of Artificial Intelligence (IIAI) from the United Arab Emirates and the University of Amsterdam. Scientific co-director of the lab Cees Snoek says “We are thrilled about this collaboration with the IIAI, a powerhouse in artificial intelligence research, and look forward to invent new technology for medical imaging together.” Cees Snoek, presented AIM Lab as the eleventh lab in the network of the national Innovation Center for AI (ICAI) and explained how such lab partnerships drive AI innovation, talent development and ecosystem growth.

Bram van Ginneken

Geert ten Dam

Victor Everhardt


After a word of welcome by Peter van Tienderen, the dean of the Faculty of Science of the UvA, Bram van Ginneken, professor at Radboud University Medical Center, gave a keynote on medical imaging. He explained its applications and the need for more research in the deep learning and computer vision technology underlying medical imaging.

This formed the bridge to the presentation of Marcel Worring, the other scientific co-director of AIM Lab. Marcel showed what the seven PhDs in the project will do on this topic and how the excellent scientific infrastructure of IIAI and UvA will help to create optimal synergy.  Before the official opening, Geert ten Dam gave a short speech indicating how AIM Lab fits in the Amsterdam agenda ‘AI Technology for People’ followed by Victor Everhardt who emphasized the important role he sees for AIM Lab and Science Park Amsterdam for the ‘Life Sciences and Health Strategy’ of Amsterdam.

More info on AIM Lab