Invited speakers
Lina Ye
Enseignante-Chercheuse, LMF, CentraleSupélec, Personal websiteLina Ye is an associate professor (MCF) of computer science at CentraleSupélec, University Paris-Saclay. Her research is on formal methods for the verification of complex systems for properties such as diagnosis-based properties (diagnosability, predictability, manifestability and so on) and safety properties. She works on infinite probabilistic models, learning systems (DNNs such as RNNs), distributed systems, discrete-event systems, real-time systems, and hybrid systems. She is part of the LeaRNNify and SAIF PEPR projects, combining formal methods with machine learning.
She received her PhD degree in Computer Science from University of Paris-Sud 11 in 2011 on optimized diagnosability of distributed discrete event systems through abstraction. She held post-doctoral positions in the research team CONVECS at Inria Rhône-Alpes and at the LRI laboratory (Laboratoire de Recherche en Informatique).
Satwik Bhattamishra
Cartesia, Personal websiteSatwik Bhattamishra is a researcher at Cartesia AI, where he works on pretraining and building LLMs with novel architectures. His research interest include the theoretical and scientific understanding of deep learning models, and more generally questions of learnability. He works on (i) Expressivity: understanding what tasks and functions architectures like Transformers or RNNs can theoretically represent; (ii) Learning theory: characterising the learnability of neural networks and automata with examples and/or query access; and (iii) Empirical learnability: understanding the algorithmic learning capabilities of neural sequence models and how we can improve them.
He recently finished his PhD in Computer Science at the University of Oxford on Expressivity and Learnability of Neural Sequence Models. Before that, he was a student researcher at Google, working on improving LLM agents; he interned at Cohere, working on pretraining LLMs with non-Transformer architectures; and he was at Microsoft Research India.
Tom F. Sterkenburg
Munich Center for Mathematical Philosophy (LMU), Personal websiteTom Sterkenburg is an Emmy Noether group leader at the Munich Center for Mathematical Philosophy (MCMP, LMU Munich), associated with the Munich Center for Machine Learning (MCML). His work is about applying the mathematical field of machine learning theory to philosophical questions around machine learning and artificial intelligence. His Emmy Noether project 'From Bias to Knowledge: The Epistemology of Machine Learning' is concerned with the fundamental notion of inductive bias.
He is also a fellow of the Konrad Zuse School of Excellence in Reliable AI (relAI) and a member of the Young Center of the LMU Center for Advanced Studies (CAS^LMU), where he co-leads a research focus on Bayesian reasoning in science. He obtained his PhD at the CWI and the Faculty of Philosophy of the University of Groningen in 2018, on the theory of universal prediction stemming from algorithmic information theory (Kolmogorov complexity).