Talk outline: We first see an overview of recent approaches to extract simpler abstractions of complex neural networks using Angluin’s exact learning framework, from computational learning theory. The aim of constructing such abstractions is to obtain high level information from machine learning models, which can be useful to interpret their behavior, detect harmful biases, among others. We then discuss in more detail algorithms for learning logical theories in Angluin’s framework, in particular, those for learning rules in Horn logic. We highlight the benefits and shortcomings of these approaches. Finally, we present promising possible next steps and applications of these approaches for extracting high level information from complex machine learning models such as large language models

Bio: Ana Ozaki is an associate professor at the University of Bergen and at the University of Oslo, Norway. She is an AI researcher in the field of knowledge representation and reasoning and in learning theory. Ozaki is interested in the formalization of the learning phenomenon so that questions involving learnability, complexity, and reducibility can be systematically investigated and understood. Her research focuses on learning logical theories formulated in description logic and related formalisms for knowledge representation. She is the principal investigator of the project Learning Description Logic Ontologies funded by RCN.

Talk outline:In recent decades, several approaches have drawn analogies between pattern mining tasks and constraint programming. The development of modern constraint solvers (SAT, CP, Linear Programming, Answer Set Programming, etc.) has demonstrated the efficiency of these approaches on real-world datasets. More than the efficiency of declarative programming, we argue that the true benefit of this approach lies in its versatility and its ability to integrate expert knowledge. This is especially evident in the case of Answer Set Programming (ASP), originally designed as a tool for knowledge reasoning. In this presentation, we will focus on the task of sequential pattern mining, which involves discovering interesting patterns in a collection of sequences. Firstly, we will introduce the concept of ASP encoding for sequential pattern mining tasks and showcase the advantages of declarative programming in fast-prototyping complex mining tasks. Next, we will highlight the capability of the Clingo solver to efficiently combine reasoning and procedural programming to address the mining of chronicles. Finally, we will conclude this presentation by exploring the potential development of epistemic measures of interestingness at the crossroad of declarative pattern mining and knowledge reasoning. To illustrate the effectiveness of a knowledge-centric approach, we will provide practical examples of these advanced features, focusing specifically on their application in analyzing care pathways to answer pharmaco-epidemiological questions.

Bio: Thomas Guyet received his Ph.D. in Computer Science in 2007 from the National Polytechnic Institute of Grenoble, France. He subsequently joined the IRISA laboratory as an Assistant Professor until 2020. His research interests lie in the analysis of temporal data, particularly time sequences, using various artificial intelligence techniques, ranging from machine learning to ontological reasoning and pattern mining. Thomas Guyet is currently a full researcher at the Inria Center of Lyon. He is a member of the AIstroSight team, which focuses on leveraging computer science tools to enhance drug design, with a particular emphasis on combining modeling and machine learning. He develops different approaches to extract knowledge from healthcare pathways and is currently a co-holder of an Inria/APHP chair in AI-Raclès dedicated to this subject.

Talk outline: TBD

Bio: Nick Chater is Professor of Behavioural Science at Warwick Business School. His research focuses on the cognitive and social foundations of rationality, with applications to business and public policy. He has (co-)written more than 250 research papers and eight books. His book, The Mind is Flat, won the American Association of Publishers PROSE Award in 2019, for Best book in Clinical Psychology. Nick is a fellow of the British Academy, the Cognitive Science Society and the Association for Psychological Science. His research has won awards including the British Psychological Society's Spearman Medal (1996); the Experimental Psychology Society Prize (1997); and the Cognitive Science Society's life-time achievement award, the David E Rumelhart Prize (2023).

Talk outline: TBD

Bio: Alan Bundy is Professor of Automated Reasoning in the School of Informatics at the University of Edinburgh. His research interests include: the automation of mathematical reasoning, with applications to reasoning about the correctness of computer software and hardware; and the automatic construction, analysis and evolution of representations of knowledge. His research combines artificial intelligence with theoretical computer science and applies this to practical problems in the development and maintenance of computing systems. He is the author of over 300 publications and has held over 60 research grants. He is a fellow of several academic societies, including the Royal Society, the Royal Society of Edinburgh, the Royal Academy of Engineering and the Association for Computing Machinery. His awards include the IJCAI Research Excellence Award (2007), the CADE Herbrand Award (2007) and a CBE (2012). He was: Edinburgh's founding Head of Informatics (1998-2001); founding Convener of UKCRC (2000-05); and a Vice President and Trustee of the British Computer Society with special responsibility for the Academy of Computing (2010-12). He was also a member of: the Hewlett-Packard Research Board (1989-91); the ITEC Foresight Panel (1994-96); both the 2001 and 2008 Computer Science RAE panels (1999-2001, 2005-8); and the Scottish Science Advisory Council (2008-12).

Talk outline: Not too long ago most headlines talked about our fear of AI. Today, AI is ubiquitous, and the conversation has moved on from whether we should use AI to how we can make the AI systems that we use in our daily lives trustworthy. In this talk I look at some key technical ingredients that help us build confidence and trust in using intelligent technology. I argue that intuitiveness, adaptability, explainability and inclusion of human domain knowledge are essential in building this trust. I present some of the techniques and methods we are building for making AI systems that think and interact with humans in more insightful and personalised ways, enabling humans to better understand the solutions produced by machines, and enabling machines to incorporate human domain knowledge in their reasoning and learning processes.

Bio: Mateja Jamnik is developing AI techniques for human-like computing – she models how people solve problems to enable machines to reason and learn in a similar way to humans. She applies AI and reasoning techniques to medical data to advance personalised cancer medicine, and to education to personalise tutoring systems. Mateja is passionate about bringing science closer to the public and engages frequently with the media and outreach events. She has been advising the UK government on policy direction in relation to the impact of AI on society.

Talk outline: TBD

Bio: Antonio Lieto is Associate Professor in Computer Science at University of Salerno (Italy) and a researcher at the ICAR-CNR Institute in Palermo (Italy). His main research topics include commonsense reasoning, language and knowledge technologies, cognitive architectures for intelligent interactive agents (embodied and not). He has been Vice-President of the Italian Association of Cognitive Sciences (AISC, 2017-2022), the recipient of the “Outstanding BICA Research Award” from the Biologically Inspired Cognitive Architecture Society (USA), and is an ACM Distinguished Speaker on the topics of cognitively inspired AI. He has authored the book “Cognitive Design for Artificial Minds” (Routledge/Taylor & Francis, 2021).

Talk outline: The ‘AI alignment problem’ refers to the possibility that AI systems may develop goals that are in conflict with those of their human creators. Part of the problem, I will argue, stems from the fact that computational theories of cognitive science lack a viable account of goal-based behavior in humans and other organisms. Without understanding what our goals are, where they come from, and how we align them with one another, there is little hope of designing AI systems that align with our goals. I argue that non-representationalist, “4E” theories of cognitive science can shed some light on this issue. These theories emphasize that biological intelligence is, at its core, about the dynamic coordination of brain, body, and environment. I will present some of my own work on trying to model the mechanisms of coordination at multiple scales of analysis, and propose that this line of work may help us design AI systems that spontaneously align with the world around them, with each other, and with their human users.

Bio: Ben Falandays is an Assistant Professor in the School of Social and Behavioral Sciences at Arizona State University. His research program focuses on building computational models that bridge multiple domains within cognitive science, including neuroscience, development, perception-action, social interaction, language, and culture.

Talk outline: The ability to learn programs from a few examples allows users to automate repetitive tasks in an intuitive way. Existing frameworks on inductive synthesis only perform syntactic manipulations, where they rely on the syntactic structure of the given examples and not their meaning. Any semantic manipulations, such as transforming dates, have to be manually encoded by the designer of the inductive programming framework. In this talk, I will highlight two methods of integrating semantic information with programming by example. First, we extract properties about the input public from public knowledge graphs and provide these properties as inputs to the model. Second, we use the in-context learning capabilities of large language models to solve small subproblems that the deductive backpropagation algorithm behind FlashFill cannot solve. I will conclude with a comparison of advantages and disadvantages of both approaches.

Bio: Gust Verbruggen is a researcher at Microsoft. His broad research interest is code generation from different modalities, such as examples, demonstrations, natural language, or combinations of them. He recently completed his Ph.D. on program synthesis for data wrangling at KU Leuven.