[FM-India]Post-doc hiring in Singapore-French collaborative program Descartes on Formal Reasoning and AI

*Post-doc hiring in Singapore-French collaborative program Descartes on Formal Reasoning and AI* Several Post-doctoral positions are invited in the 5-year research program Descartes: Intelligent Modeling for Decision-making in Critical Urban Systems. This program is jointly between National University of Singapore (NUS), CNRS France, and several organizations in Singapore and France. One facet of the program deals with formal reasoning and AI, for which we are seeking to hire a postdoctoral research fellow. The research work is on trustworthy AI, intersecting between formal methods and AI. See https://descartes.cnrsatcreate.cnrs.fr/wp-1-trustworthy-hybrid-ai/ It will focus on one or more of the following research themes: 1.We want to provide /formal /guarantees about the intelligent systems, considering all possible behaviors. In doing so, combining formal reasoning with learning processes, we will contribute to the development of /neuro-symbolic reasoning/. Moving forward from verification, it would also be critical for autonomous systems to employ self-healing, subject to adaptations in the environment which includes humans. In this context, there exist significant opportunities in /automated code generation and rectification/ using language models, and using formal reasoning to enhance trust in the generated code. 2.At the same time, the verification of hyper-properties of intelligent systems has gained importance, with newer techniques for privacy, robustness and fairness. Formal methods to establish hyper-properties in intelligent decision-making may thus decompose a high-dimensional input space into hyper-cubes and calculate the “desired” probabilities of the hyper-cubes to preserve an overall hyper-property for the entire system. Alternatively, the probabilistic nature of hyper-properties may be exploited to establish their validity by adaptively sampling inputs from a high dimensional space. 3.We will develop deductive reasoning-based explanations of hyper-property violations (apart from explaining the AI model predictions), which can then be used by neuro-symbolic reasoning technique to rectify the AI models so that it can satisfy fairness/robustness etc. We also plan to study interactive mechanisms for producing explanations while considering the cognitive load of understanding the explanations by human users. The post-doc will work from CREATE Tower on the National University of Singapore campus which is a leading university in Asia and provides significant opportunities for career growth, supervised jointly by French and Singaporean colleagues. Interested candidates can directly contact the following Lead Principal Investigators (Lead PIs): Blaise Genest, CNRS Senior Researcher, https://perso.crans.org/genest/ Abhik Roychoudhury, Provost’s Chair Professor, National University of Singapore, https://abhikrc.com/ -- -Abhik Roychoudhury School of Computing National University of Singapore https://abhikrc.com