There are currently no vacancies.
The PhD position posted below will open later this year.
Decision Making and Law: How do professionals from different fields interact, each with their own expertise and tasks, in relation to AI systems and human decision making?
Background of the PhD student: Public administration or Law, with strong skills in applied AI, or Computer Science/AI with a strong affinity with Public administration and Law.
Institute of employment: Institute for Public Administration
The successful applicant should be a very motivated university graduate who is a top performer among his/her peers, has an excellent education and/or research track record proven by relevant experience, publications, etc. You are expected to conduct original competitive scientific research, publishing the results in top conferences and scientific journals, and participating in teaching duties.
Applicants are expected to have a university degree (MSc) in a relevant field (see the description of the positions for more details). Collaborative research across domains is of key importance in the SAILS program. You are expected to team up with your colleagues and have an open mind in discovering new and innovative relations between the different fields involved.
• Conduct original and novel research in the field of AI
• Publish and present scientific articles at international journals and conferences;
• Contribute to educational activities;
• Write a dissertation.
• Engage in collaborations with colleagues across domains
• A MSc degree in a relevant field (see descriptions);
• Good knowledge of and experienced with AI techniques;
• Highly motivated to both perform foundational research and apply the developed methods to real-world problems;
• Creative, ‘making things work’ mentality, independent, and communicative team player;
• Experienced with writing scientific manuscripts and good academic writing skills;
• Excellent programming skills (preferably C++ and/or Python);
• Interested in participating in educational activities;
• Excellent proficiency in English (oral and written).
Terms and conditions
We offer a full-time 1 year term position for initially one year. After a positive evaluation of the progress of the thesis, personal capabilities and compatibility the appointment will be extended by a further three years. Salary range from € 2.325,- to €2.972,- gross per month (pay scale P in accordance with the Collective Labour Agreement for Dutch Universities).
Leiden University offers an attractive benefits package with additional holiday (8%) and end-of-year bonuses (8.3 %), training and career development and sabbatical leave. Our individual choices model gives you some freedom to assemble your own set of terms and conditions. Candidates from outside the Netherlands may be eligible for a substantial tax break.
All our PhD students are embedded in the Leiden University Graduate School. Our graduate school offers several PhD training courses at three levels: professional courses, skills training and personal effectiveness.
Leiden University is strongly committed to diversity within its community and especially welcomes applications from members of underrepresented groups. We wish to reflect society both in age, gender and culture, as we believe that this would optimize the dynamics in our organization. Therefore, we support and understand the need for a work/life/family balance and consequent varying working hours and places. In the Netherlands, a maternity allowance is standard for 16 weeks. Child care is available at and near the Bio Science Park.
Enquiries can be made to the SAILS Coordinator
To apply for this vacancy, please send an email to SAILS Coordinator
When applying, please use the following subject: “SAILS – Your name”. Please ensure that you attach the following additional documents, quoting the vacancy number:
• Curriculum vitae (CV);
• Motivation letter (1 page maximum);
• Grade list and MSc degree (or expected graduation date);
• (Draft of) MSc thesis;
• Two reference letters (at least one from your Master thesis supervisor);
• Link to public code repository (e.g., GitHub) or example of written code base.
Only applications received no later than TBD can be considered. Applicants may be contacted with requests for additional information and/or exploratory conversations before the closing date.
The city of Leiden
Leiden University is one of Europe's foremost research universities. Leiden is a lively and beautiful city with a central location in Europe, at only 17 minutes from Schiphol International Airport. It is a very international city where almost everybody speaks English, candidates need not be afraid of the language barrier. Leiden is a typical historic university city, hosting the oldest university in the Netherlands (1575). The University permeates the local surroundings; University premises are scattered throughout the city, and the students who live and study in Leiden give the city its relaxed yet vibrant atmosphere. Today, with some 27,000 students, 5,500 staff, 16 Nobel prizes, more than 100 nationalities and a budget of more than 550 million Euros, it is one of the largest comprehensive universities in Europe. This prominent position gives our graduates a leading edge in applying for academic posts and for functions outside academia.
Leiden University is strongly committed to diversity within its community and especially welcomes applications from members of underrepresented groups. For more information: Diversity
LIACS strives for excellence in a caring Institute. We wish to reflect society in all aspects, as we believe that this improves the dynamics in our organization. Therefore, we support and understand the need for a work/life/family balance and consequently varying working hours and places. We believe that doing good science is fun, and goes hand in hand with a driven yet friendly culture. In the Netherlands, parental leave is standard for 16 weeks. Child care is available at and near the Bio Science Park.
A wealth of expertise in the fields of philosophy, (art) history, the arts, literature, film and cultural sciences, religious studies, linguistics and multidisciplinary area studies results in a unique Faculty of Humanities.
Our global expertise, international focus and synergy of scientific approaches form windows to the world, opening up the many perspectives of a globalising world. In a century of globalisation, large-scale levels of migration, intense and ever faster communication, cultural transformations and the politicisation of culture and the rise of new superpowers, the work and knowledge of the faculty’s staff is of vital importance to our society, both inside and outside of the Netherlands. It is precisely in the 21st century that our faculty consciously stands at the heart of that society.
The Faculty of Humanities offers an inspiring and international work environment with room for diversity and reform for employees and students from the Netherlands and abroad. It offers a wide range of education, with students being able to choose from no less than 26 Bachelor’s and 27 Master’s degree programmes. The Faculty’s research is organised within seven institutes and contributes to six research dossiers for research at Leiden University.
Faculty of Governance and Global Affairs
The Institute of Public Administration is one of the largest and oldest institutes of academic research and teaching in the field of public administration in the Netherlands and is located in The Hague. The institute combines a solid international academic reputation with a central positioning among the international, national, regional and local governance institutions of The Hague. It employs around 80 people and provides education to 1200 students. It is part of the Faculty Governance and Global Affairs (FGGA), one of the seven faculties of Leiden University. Besides Public Administration the faculty consists of two other institutes and four centers of expertise. More information about the faculty may be found at: Faculty of Governance and Global Affairs
The institute has an international profile and gets high peer reviewed ratings, both for its research and education. It offers a Dutch Bachelor of Public Administration, a Dutch Master Management of the Public Sector and an English Master Public Administration. All three have several specializations. Based on the strong academic disciplinary approach, the institute collaborates with other institutes and faculties of Leiden University on interdisciplinary research and education programs.
With over 5,000 students and 450 members of staff, Leiden Law School is one of the largest faculties in the Netherlands. Yet, in all its diversity, it is still known for its ability to provide education on a small scale. The Faculty focuses on multi-faceted high-level teaching and research, both nationally and internationally. It does so by working with talented people and stimulating and supporting them in their professional and personal ambitions. The Faculty is housed in the beautifully restored Kamerlingh Onnes Building on the Steenschuur in Leiden. Working for the Leiden Law School means working in an inspiring scientific environment. For more information, see: Law.
The center for law and digital technologies (eLaw) was founded in 1985 and has a leading role in research and education on law and digital technologies. The center studies the social, legal and normative impact of emerging digital technologies. The focus of our research and education is particularly on digital technologies and their interplay with fundamental rights and the rule of law. Key themes in our research and education are: online privacy and personal data protection; cybercrime and cybersecurity; data science and law; internet governance; law and complexity in digital society; electronic communications law; children's rights in the digital world; and media and fundamental rights. Information about the center for law and digital technologies can be found at: eLaw and about Leiden University at Education.
Faculty of Social and Behavioural Sciences / Institute of Psychology
The Faculty of Social and Behavioural Sciences comprises four institutes: Education and Child Studies, Political Science, Psychology and Cultural Anthropology & Development Sociology. The Faculty also includes the Centre for Science and Technology Studies and hosts 5,000 students and 600 staff members.
The Institute of Psychology comprises six units (teaching + research): Health, Medical and Neuropsychology, Clinical Psychology, Cognitive Psychology, Developmental and Educational Psychology, Social and Organizational Psychology, and Methodology and Statistics. The Institute of Psychology offers a stimulating environment that promotes collaboration within and between units. For more information, see Psychology.
The institute has a strong international orientation. Its mission is to perform high quality research at the frontiers of mathematical knowledge, and to educate future generations of mathematicians and statisticians in a friendly but challenging environment.
The expertise of the members of the institute covers a broad range of topics, with a focus on Algebra, Geometry and Number Theory; Analysis and Dynamical Systems; Probability Theory; and Statistics. Research is focused both on fundamental mathematics and statistics and on applications in other sciences, society and industry.
Leiden Academic Centre for Drug Research
The division of BioTherapeutics is part of the Leiden Academic Centre for Drug Research (LACDR). Our institute is a centre of excellence for multidisciplinary research on drug discovery and development. Leiden University consistently ranks within the top-20 world-wide within the subject-based QS Ranking in Pharmacology and Pharmacy. At LACDR, we develop innovative scientific concepts in fundamental and translational research underlying the discovery and optimization of drugs, and personalised medicines. We aim to optimise the route to novel drugs that are both efficacious and safe, and therefore our research focuses on drug and target discovery, bio-therapeutics and systems pharmacology. State-of-the-art expertise and infrastructures ensure that we are strategically positioned in (inter)national collaborations, thus further strengthening the development of novel concepts and their application in a clinical and industry environment.
LACDR has close on-going collaborations with the Leiden University Medical Centre (LUMC) and the Centre for Human Drug Research (CHDR). More information about the LACDR can be found at: LACDR.
The Leiden Institute of Chemistry
The Leiden Institute of Chemistry (LIC) is the basis for research and collaborations of the Leiden chemistry groups. Chemistry is the central science enabling a healthy future in a sustainable society. Chemistry researchers at Leiden University take a fundamental approach in finding tailored solutions for complex societal problems in human health and environmental issues.
The chemistry and life science research in the Leiden Institute of Chemistry is organised around two major research clusters: Chemical Biology and Energy & Sustainability.
Leiden Institute of Advanced Computer Science
The Leiden Institute of Advanced Computer Science (LIACS) is a centre of excellence for multidisciplinary research and education in computer science and artificial intelligence (AI).
It is our mission to improve current computer science methods, systems and techniques, whilst exploring new research areas that are relevant to society. We concentrate on the study of theoretical foundations and formal methods, and focus on applications in the field of artificial intelligence and data science.
With (big) data as a driving force in all fields of science, LIACS has a multidisciplinary perspective striving for research results that combine computer science and other disciplines at Leiden University.
LIACS is known for its strong integration of research and education. Our international environment creates many possibilities for research projects and offers a close collaboration with research groups.
Research at the Faculty of Science
The Faculty of Science is a world-class faculty where staff and students work together in an international environment. It is a faculty where personal and academic development are top priorities. Our people are driven by curiosity to expand fundamental knowledge and to look beyond the borders of their own discipline; their aim is to benefit science, and to make a contribution to addressing the major societal challenges of the future.
The research carried out at the Faculty of Science is very diverse, ranging from mathematics, information science, astronomy, physics, chemistry and bio-pharmaceutical sciences to biology and environmental sciences. The research activities are organised in eight institutes. These institutes offer eight bachelor’s and fourteen master’s programmes. The faculty has grown strongly in recent years and now has more than 1,300 staff and over 4,000 students. We are located at the heart of Leiden’s Bio Science Park, one of Europe’s biggest science parks, where most companies have its focus on drug research. For more information, see Science and Working at Leiden University
1. SAILS PhD project on Imaging: Deep-Learning-Based Depth Recovery from 2D Images
Project lead: Joost Batenburg (LIACS/FWN)
Institutes: 1. LKEB/LUMC; 2. Archaeology; 3. LIACS/FWN; 4. MI/FWN; 5. NeCEN/IBL
In recent years, 3D object extraction from monocular imagery has increasingly became a major research interest, especially thanks to the investments in autonomous driving. While the under-constrained inverse 2D-to-3D geometric problem has been an intellectual challenge, recent advancements in computer technologies opened up a new research avenue. In particular, there is a growing research in deep-learning with the aim of increasing the accuracy of depth information retrieval from single or multiple complementary views of the same scene and defining proper geometric constraints for 3D object detection. However, despite the recent boost obtained by this subject with rapid advance of deep-learning methodology, its practical applications are primarily limited to reconstruction of single objects. Availability of the training data is the main bottleneck that renders wide application of this methodology in different areas, in particular for parsing complex outdoor scenes, very challenging. Hence, weakly- and semi-supervised methods in combination with other sources of information are currently seen as a promising research direction.
In this project, we will develop novel methodology for recovery of depth information from 2D projection data. Potential research directions include:
• Developing neural-network architecture that would rely on geometric and photometric ques (shading, perspective) for 3D scene reconstruction (bottom-up approach).
• Efficient (semi-supervised) strategy for depth annotation based on available data.
• Development of reliable projection methods for generation of synthetic data for training.
• Adding complementary class-specific (e.g. material) information for simultaneous object classification (top-down approach); possibility to combine it with the described bottom-up approach.
Foundational research question
How can we construct a deep-learning-based methodology for 3D scene reconstruction from projection images with broad application in various disciplines.
The methodology will be implemented into a working prototype, with a separate user interface for each application.
At the initial stage of this project, existing state-of-the-art reconstruction methodology will be applied for improvement of depth resolution of 3D scanning electron microscopy (SEM) data. Consequently, this methodology will be extended for solving more challenging problems, such as reconstruction of complex scenes consisting of multiple “unseen” objects, under natural illumination conditions, etc. The ultimate goal of this project is to enable reconstruction of multi-object outdoor scenes from historical photographs of archaeological excavations.
Team and embedding
LUMC: Machine learning, Human computer interaction, Microbiological application
LIACS: Machine learning, Human computer interaction
Archaeology: Remote sensing, Archaeology application
MI: dynamical systems
NeCEN: Electron microscopy application
2. SAILS PhD project on AI and Humanities: Socially embedded AI systems for natural machine learning and human-machine interaction.
Project lead: Tessa Verhoef (LIACS/FWN)
Institutes: 1. LIACS/FWN; 2. Cognitive Psychology/FSW; 3. LUCL/HUM
AI involves the creation of systems that can perform behaviors for which human intelligence is needed. More and more, the kinds of cognitive abilities AI researchers are trying to teach machines are associated with complex social behaviors such as language, music, dance, art, technology and emotional expression. Currently, AI engineers expect individual robots or agents to learn these complex activities entirely on their own, through exposure to large amounts of data. But humans are community animals, deeply ingrained and supported by culture. A lot of complex human behaviors evolve culturally and couldn't exist without a community of brains using and transmitting the systems that support these behaviors. For robots and machines to become effective and successful members of our society, they need to gradually and flexibly adapt their behavior and communicatively co-evolve with us.
In this PhD project, adaptive approaches for human-machine communication will be explored with the following aims:
1. to enable more natural interactions between humans and AI systems
2. to create more natural settings for machine learning algorithms to learn complex human-like behaviors.
Candidate adaptive approaches include (deep) reinforcement learning, evolutionary computing and, possibly, combinations thereof. The project focuses on the emergence of natural human-robot communication starting from simple synchronizing behaviours, a novel approach to natural language processing and AI experiments with a strong focus on human social cognition, reinforcement learning and self-organization and cultural transmission. Outcomes of the research could shed light on the relationship between symbolic (language) and sensorimotor representation and communication. An additional focus of the PhD project will be natural language dialogues for the purpose of explainability: how can AI models learn from and interact with humans in natural dialogues aimed at conveying explanations? To this end, representative observational data will be gathered (human explanation dialogues).
Foundational research question
How can we model the emergence of natural human-machine interaction and allow AI systems to acquire complex communicative behavior, for example in the context of natural dialogues, and by studying the affective component of reward in reinforcement learning tasks and more fundamental social coordination tasks.
Software / demonstrator
Software/output will consist of a suite of AI models, demonstrations and code notebooks. One of the additional outcomes is a dataset with transcribed human dialogues. This dataset will be a valuable resource to study communicative patterns between humans, and will be used as benchmark data for the AI approaches to be developed.
The PhD candidate will apply the research to the development of Natural Language Processing systems, leveraging (and developing) novel approaches in reinforcement learning and evolutionary computing and insights into language evolution. The knowledge acquired will be valuable for HCI, AI- and LUCL (linguistics, cognition) researchers
Team and embedding
LUCL: NLP expertise, semantic models, Deep Learning expertise (deep NLP, reinforcement learning), application
FSW: Cognitive Psychology expertise, Evolutionary Robotics expertise, Neural Computing
LIACS: Deep Learning expertise, Human-machine interaction expertise, Cultural Evolution, Social Cognition
3. SAILS PhD project on Drug Discovery: Exploration of Bioactivity Space with Artificial Intelligence
Project lead: Gerard van Westen (LACDR/FWN)
Institutes: 1. LACDR/FWN; 2. LIC/FWN; 3. LIACS/FWN
Data science and artificial intelligence have emerged as major players in the life sciences. Using machine learning, molecules can be generated in silico, allowing the extremely high throughput exploration of this massive chemical space. Subsequently, using machine learning, these generated molecules can be mined for new drugs for existing diseases. However, available bioactivity data of existing molecules is orders of magnitude lower (around 106 have been characterized), very sparse, and historically biased towards a limited number of well explored protein drug targets. Moreover, the massive throughput in computational methods cannot be matched by biological experiments. Hence, there is a significant gap between computational capacity, the available bioactivity information in literature, and the potential chemical structures available for drug discovery (“Dark Bioactivity Space”).
In this PhD project we will use artificial intelligence (machine learning) to rationally and iteratively explore the prediction limits of publicly available data (starting with the ChEMBL database). This project will consist of two steps. Step one focuses on two relevant protein targets in drug discovery (the cannabinoid receptors). Step two consists of scaling successful tools up to cover all relevant proteins (i.e. with enough ligands) in the entire ChEMBL database and see how our predictions can assist ongoing research at Leiden University. In both phases we will identify missing information about the biological activity of small molecules using active learning.
Our active learning approach will consist of selecting the data points that are most information rich to the training set for a machine learning model. Determining what is most information rich will be done in two ways via ‘curious’ active learning. In the ‘curious’ approach datapoints will be added that are predicted to be highly active, but where the probability is close to random (i.e. on the border of the decision boundary). To validate this retrospectively we will use the ‘greedy’ approach wherein data points are added that are predicted to be active but have a very high probability (i.e. the model is certain). We hypothesize that the curious approach will lead to a quick increase in data set quality.
Secondary objectives are the identification of novel selective and promiscuous tool compounds to explore the bioactivity space. Moreover, the creation and distribution of predictive bioactivity models that can be readily added to existing workflows will be included in the project (and can be implemented by both academic and commercial partners). Finally, using multiple prospective validation experiments, the prediction quality of machine learning models will be explored and the effect of ensembling methods on this reliability will be quantified.
Foundational research question
How can we use machine learning and optimization techniques for predicting which molecules should be chosen for further experimental investigation.
Software / demonstrator
The PhD project will result in a software demonstrator based on the ChEMBL database.
The key application is high-throughput chemical space exploration for drug research. Based on the ChEMBL database, we will search for molecules that target the cannabinoid receptors and then scale up to cover all relevant proteins (i.e. with enough ligands) in the entire ChEMBL database.
Team and embedding
LACDR: Machine learning, Drug research application
LIACS: Machine learning, Reinforcement learning, Optimization
4. SAILS PhD project on Decision making and Law: Understanding and improving the alignment between AI derived insights and decision-making by professionals
Project lead: Bram Klievink
Institutes: 1. Public Administration/FGGA; 2. e-Law/LAW; 3. LIACS/FWN
Various forms of AI are developing rapidly and are increasingly being used in domain-specific contexts, such as by medical professionals, businesses executives, administrative decision-makers, and policy makers. There is much enthusiasm as to the functional benefits that AI might bring to human decision makers. However, it is notoriously hard to understand and explain how algorithms work and thus fully assess how they contribute to the intelligence and insights that can be used in or underly decision-making. Consequently, various choices (including methodological or technical), limitations and trade-offs reflected in various steps in the chain towards intelligence may be lost on the user. On the other hand, the developers of the AI-system may not be aware of specific sensitivities of the context of use that could inform choices in the trade-offs.
For example, an administrative professional with plenty of domain expertise, but limited technical or statistical expertise, may rely too much on the outcomes of a classifier due to a limited understanding of the limitations posed by the algorithms used, the data fed into it, or the presentation of the outcomes. Or a medical doctor, facing a false-positive medical diagnoses repeatedly, may lose trust in the system, setting back the actual uptake significantly, also if it was mostly a matter of requiring better training data for a specific type of patients that is specific to their hospital. Or a policy maker, having to decide on a policy on applying facial recognition to set up a digital perimeter in high-risk areas, may not grasp how facial recognition works, and how to interpret its consequences.
Yet, there is also the other side; the provision of AI derived insights. The people mentioned above use AI as part of making (small or big) decisions on a topic they have great professional expertise on; they have best-in-class knowledge about specific policy areas, medical conditions, patients, or their branch of businesses. They are professionals. Those working on the supply side of AI, are also professionals, although with a completely different expertise; technical, analytical and methodological rather than domain-specific, medical or political. In this project we study the interaction between these two worlds: the world of AI and that of human decision making.
Foundational research questions
1. Understanding: What decisions are made along the chain of development and use of algorithmic decision systems, how and by whom, how are they reported on, and how do they affect an eventual policy or decision?
2. Improving: which interventions (technical, organisational, regulatory) can and should be developed to enable optimal use of AI in decision making by professionals?
Software / demonstrator
As part of this PhD project, a flexible and adaptable demonstrator of an algorithmic decision system will be developed to facilitate the analysis and allow for experimenting with different scenarios.
The application field concerns the broad set of algorithmic decision systems where the knowledge of domain experts and AI experts must be combined to achieve reliable and effective results.
Team and embedding
Public Administration/FGGA: Decision making, public policy
eLaw/LAW: Legal aspects of AI systems
LIACS: Modelling human behaviour