AILA 2025

Keynote Speeches

Title: Logic for Reinforcement Learning

Natasha Alechina
Open University of Netherlands and Flanders

Bio: Natasha Alechina is a Professor of Safe and Responsible AI at the Open University of Netherlands and Flanders. She also holds a part-time position at Utrecht University. She obtained her undergraduate degree in logic from Moscow State University and her PhD from the Department of Mathematics and Theoretical Informatics of the University of Amsterdam. She worked at Computer Science department at the University of Birmingham and at the University of Nottingham in the UK, working on applications of logic in computer science and artificial intelligence. Her current research spans several areas of Artificial Intelligence, including: verification of multi-agent systems for quantitative properties (such as resource requirements), synthesis of multi-agent programs, synthesis of norms, causal reasoning, assigning responsibility in teams of human and AI agents. She has published over 100 papers in top international venues.

Abstract: Temporal logic is increasingly used in Reinforcement Learning to give a declarative specification of a task to be learned (and avoid problems with reward gaming and other challenges of reward functions) and to specify declarative constraints on agents’ behaviour (in particular, safety constraints). I will talk about recent work in this area.

Title: Extending Hoare logic to hybrid systems

Naijun Zhan
Peking University, School of Computer Science

Bio: Naijun Zhan is a Boya distinguished professor in the School of Computer Science of Peking University. He got his bachelor’s degree and master’s degree both from Nanjing University, and his PhD from Institute of Software Chinese Academy of Sciences (ISCAS). Prior to join Peking University, he worked at the Faculty of Mathematics and Informatics, Mannheim University, Germany as a research fellow, and afterwards worked at ISCAS as an associate professor, a full professor, and a distinguished professor. His research interests cover formal design of real-time, embedded and hybrid systems, program verification, and so on. He is in the editorial boards of Journal of Automated Reasoning, Formal Aspects of Computing, Journal of Logical and Algebraic Methods in Programming, Journal of Software, Journal of Electronics, and Journal of Computer Research and Development and so on, a member of the steering committees of SETTA and MEMOCODE, the pc co-chairs of FM 2021, SETTA 2016, the general co-chairs of MEMOCODE 2018, MEMOCODE2019 and ICESS 2019, and serves more than 100 international conferences program committees e.g., CAV, RTSS, HSCC, FM, TACAS, EMSOFT and so on. He published more than 150 papers in leading international journals and conferences and 2 books, and edited 4 conference proceedings and 7 journal special issues. See lcs.ios.ac.cn/~znj for more details.

Abstract: In this talk, I will talk our efforts on extending Hoare logic to hybrid systems, named hybrid Hoare logic, including two versions: a DC-based version and and a generalized version without DC. I will briefly introduce the DC-based version, and mainly focus on the second version. I will explain why we need a generalized version, and present its proof system, the soundness and discrete and continuous relative completeness of the proof system, the implementation of an interactive theorem prover based on Isabelle/HLL called HHLProver, the automation of HHLProver, and finally introduce some applications with the logic and its prover.

Title: Ensemble Logic: a Unified Framework for Temporal, Spatial, and Spatial-Temporal Reasoning

Guo-Qiang Zhang
The University of Texas Health Science Center at Houston

Bio: Guo-Qiang Zhang is distinguished chair in digital innovation, vice president and chief data scientist at the University of Texas Health Science Center at Houston (UTHealth Houston). He serves as co-director for the Texas Institute for Restorative Neurotechnologies and holds primary appointment as full professor in the Department of Neurology at UTHealth Houston's McGovern Medical School. He has secondary joint appointments in the McWilliams School of Biomedical Informatics and School of Public Health, UTHealth Houston.
Before joining UTHealth, he served as the inaugural director for the Institute for Biomedical Informatics, chief of the Division of Biomedical Informatics, and associate director for the Center for Clinical and Translational Science (CTSA) at the University of Kentucky. He spent prior years as a faculty member at the Case School of Engineering and School of Medicine, at Case Western Reserve University, where he created its Division of Biomedical Informatics in the School of Medicine while serving as its CTSA informatics core director.
GQ Zhang received his Ph.D. in Computer Science from Cambridge University. His research work and interests span data ecosystems and big data, biomedical ontology development and quality assurance, clinical and research informatics, and data coordination systems for prospective metadata management in national consortia, such as the latest, on-going NIH BRAIN Initiative Cell Atlas Network (BICAN). He has launched a research program called Logic-based Phenotyping, a formalized mathematical approach for temporal phenotype representation and reasoning using an expressive and purpose-fitting framework called Temporal Ensemble Logic. His work has been funded by research awards from the National Institutes of Health (NIH) and the National Science Foundation (NSF) of the United States.

Abstract: I introduced Temporal ensemble logic (TEL; arXiv:2408.14443 ) a year ago to fill a gap in formal languages for capturing phenotypes in biomedicine. TEL combines first-order logic and modal logic constructs as primitives in a novel setup that is more expressive than existing linear-time temporal logic. In this presentation I will cover the latest development in this area, with motivating examples and expressiveness results in terms of language classes and undecidability properties, and introduce Ensemble Logic (EL) as a generalized, unified framework for (1) temporal, (2) spatial, and (3) spatial-temporal reasoning on any mathematical space consisting of a monoid with a distance measure or an order relation, such as real numbers ( ℝ ), integers ( ℕ ), and Euclidian spaces ( ℝⁿ ).

Title: A Formal Semantics for Generic Sentences and Metaphorical Sentences

Beihai Zhou
School of Philosophy, Shanxi University

Bio: Zhou Beihai is a professor at the School of Philosophy of Shanxi University and a doctoral supervisor, as well as a professor at the Department of Philosophy of Peking University. He has served as the Director of the Logic Teaching and Research Office at the Department of Philosophy of Peking University, President of the Beijing Logic Society, and Vice President of the Chinese Logic Society. He has long been engaged in teaching and research in the field of logic, with main research areas including symbolic logic, formal semantics, logic and cognition, and logic and philosophy. He places special emphasis on reasoning issues (such as entailment, relevant logical reasoning, non-monotonic reasoning, generic sentence reasoning, etc.) and interdisciplinary research in logic, language, and cognition. He has presided over multiple projects funded by the National Social Science Fund (including major projects) and the Ministry of Education. Recently, he was awarded the 2024 National Social Science Fund Major Project “Research on the Logic of Uncertain Reasoning and Its Applications.”

Abstract: Generic sentences are sentences such as “Birds can fly”. Generic sentences are sentences that express the characteristic attributes of things. They are universal and at the same time tolerate exceptions. This characteristic is one of the fundamental reasons for the non - monotonicity of reasoning. Therefore, the logic of generic sentences is also a non - monotonic logic. The formal characterization of the reasoning of generic sentences was first proposed by default logic. There are various different understandings of generic sentences, and multiple formal characterization schemes have also emerged. These understandings and characterizations all have their own problems. This study argues that generic sentences are sentences that express the connotations of everyday concepts, and based on this, a formal semantics for generic sentences is provided. The formal semantics of generic sentences has two aspects of significance: on the one hand, it is an intermediary connecting intuitive analysis and formal characterization. Starting from the basic understanding of generic sentences, through this intermediary, one can then move on to the formal characterization of the logic of generic sentences; on the other hand, there is a natural connection between generic sentences and metaphorical sentences. If a formal semantics correctly characterizes generic sentences, then it should be able to be extended to metaphorical sentences. It is precisely based on this view that the formal semantics of generic sentences provided in this study is then extended to metaphorical sentences. The success of this extension in turn confirms the rationality of understanding generic sentences based on everyday concepts.