<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><h2 class=""><span lang="EN-US" class="" style="font-size: 14px;">PhD position in Epistemic Reasoning and Multi-Agent Learning<o:p class=""></o:p></span></h2><h2 class=""><span class="" style="font-size: 14px;">Institut de Recherche en Informatique de Toulouse (IRIT)<o:p class=""></o:p></span></h2><h2 class=""><span lang="EN-US" class="" style="font-size: 14px;">Toulouse University<o:p class=""></o:p></span></h2><h2 class=""><span lang="EN-US" class="" style="font-size: 14px;">France<o:p class=""></o:p></span></h2><p class="MsoNormal"><o:p class=""> </o:p></p><p class="MsoNormal"><span class="" style="font-family: Times; text-align: justify;">The proposed PhD thesis will be developed in context of the ANR project entitled “Learning through epistemic reinforcement” (EpiRL) which was accepted in July 2022 and will be carried out between 2023 and 2027. </span></p><p class="MsoNormal" style="margin-bottom: 12pt; text-align: justify; line-height: 15pt;"><span lang="EN-US" class="" style="font-family: Times;">The PhD thesis will start in September 2023 and will be funded on a three-year contract </span><span lang="EN-US" class="" style="font-family: Times;">with gross salary of approximately 2000€ per month. </span><span class="" style="font-family: Times;"><o:p class=""></o:p></span></p><h3 class=""><span lang="EN-US" class="">Description of the research project<o:p class=""></o:p></span></h3><p class="MsoNormal" style="text-align: justify;"><span lang="EN-US" class="" style="font-family: Times;">The need for an integration of machine learning (ML) and knowledge representation (KR) has been largely emphasized in the artificial intelligence (AI) community. </span></p><p class="MsoNormal" style="text-align: justify;"><span lang="EN-US" class="" style="font-family: Times;">According to (Valiant, 2003), a key challenge for computer science is to come up with an integration of the two most fundamental phenomena of intelligence, namely, the ability to learn from experience and </span></p><p class="MsoNormal" style="text-align: justify;"><span lang="EN-US" class="" style="font-family: Times;">the ability to reason from what has been learned. </span><span lang="EN-US" class="" style="font-family: Times;">The </span><span lang="EN-US" class="" style="font-family: Times;">PhD thesis </span><span lang="EN-US" class="" style="font-family: Times;">will be focused on the integration of epistemic reasoning and multi-agent learning. Different solutions of integration will be explored including:<o:p class=""></o:p></span></p><p class="MsoListParagraphCxSpFirst" style="margin-left: 18pt; text-align: justify; text-indent: -18pt;"><span lang="EN-US" class="" style="font-family: Symbol;">·<span class="" style="font-size: 7pt; line-height: normal; font-family: 'Times New Roman';"> </span></span><span lang="EN-US" class="" style="font-family: Times;">how to combine an agent’s capacity to attribute beliefs to other agents and to reason strategically with the capacity to form predictions about future events and future agents’ actions based on its past experiences;<o:p class=""></o:p></span></p><p class="MsoListParagraphCxSpMiddle" style="margin-left: 18pt; text-align: justify; text-indent: -18pt;"><span lang="EN-US" class="" style="font-family: Symbol;">·<span class="" style="font-size: 7pt; line-height: normal; font-family: 'Times New Roman';"> </span></span><span lang="EN-US" class="" style="font-family: Times;">how to relate the notion of reward to mental attitudes including beliefs and desires;<o:p class=""></o:p></span></p><p class="MsoListParagraphCxSpLast" style="margin-left: 18pt; text-align: justify; text-indent: -18pt;"><span lang="EN-US" class="" style="font-family: Symbol;">·<span class="" style="font-size: 7pt; line-height: normal; font-family: 'Times New Roman';"> </span></span><span lang="EN-US" class="" style="font-family: Times;">how to include in the description of a state used in an agent’s reward function the representation of other agents’ beliefs.<o:p class=""></o:p></span></p><p class="MsoNormal" style="text-align: justify;"><span lang="EN-US" class="" style="font-family: Times;">To this aim, we plan to combine concepts and methods from epistemic logic and planning (Fagin et al., 1995; Lorini, 2020; Davila et al., 2021), theories of learning in games and </span></p><p class="MsoNormal" style="text-align: justify;"><span lang="EN-US" class="" style="font-family: Times;">multi-agent learning (Fudenberg & Levine, 1998; Tuyls & Weiss, 2012), and the epistemic theory of convention (Lewis, 1969). </span><span lang="EN-US" class="" style="font-family: Times;"> </span><span lang="EN-US" class="" style="font-family: Times;">We expect the kind of integration proposed in the context of </span></p><p class="MsoNormal" style="text-align: justify;"><span class="" style="font-family: Times;">PhD thesis to be relevant for AI applications in social robotics and human-machine interaction, given the importance of combining reasoning and learning as well as prediction and explanation for such applications.</span></p><div class="" style="margin: 0cm 0cm 0.0001pt; text-align: justify;"><u class=""><span lang="EN-US" class="" style="font-family: Times;">References<o:p class=""></o:p></span></u></div><div class="" style="margin: 0cm 0cm 0.0001pt; text-align: justify;"><span lang="EN-US" class=""><font face="Times" class="">- J. Fernandez Davila, D. Longin, E. Lorini, F. Maris (2021). A Simple Framework for Cognitive Planning. In <i class="">Proceedings of AAAI-21</i>, pp. 6331-6339. <o:p class=""></o:p></font></span></div><div class="" style="margin: 0cm 0cm 0.0001pt; text-align: justify;"><span class="" style="font-family: Times;">- R. Fagin, J. Y. Halpern, Y. Moses, and M. Vardi. </span><i class="" style="font-family: Times;">Reasoning about Knowledge</i><span class="" style="font-family: Times;">. MIT Press, Cambridge, 1995.</span></div><div class="" style="margin: 0cm 0cm 0.0001pt; text-align: justify;"><span lang="EN-US" class=""><font face="Times" class="">- D. Fudenberg, D. K. Levine. <i class="">The Theory of Learning in Games</i>. MIT Press, Cambridge, 1998. </font></span></div><div class="" style="margin: 0cm 0cm 0.0001pt; text-align: justify;"><font face="Times" class="">- D. K. Lewis. <i class="">Convention: a philosophical study</i>. Harvard University Press, Cambridge, 1969.
<span lang="EN-US" class=""><o:p class=""></o:p></span></font></div><div class="" style="margin: 0cm 0cm 0.0001pt; text-align: justify;"><span lang="EN-US" class=""><font face="Times" class="">- Lorini, E. (2020). Rethinking epistemic logic with belief bases. <i class="">Artificial Intelligence, </i>282.</font></span></div><div class="" style="margin: 0cm 0cm 0.0001pt; text-align: justify;"><span class="" style="font-family: Times; background-color: white;">- K. Tuyls and G. Weiss. Multiagent Learning: Basics, Challenges, and Prospects. </span><i class="" style="font-family: Times;">AI Magazine</i><span class="" style="font-family: Times; background-color: white;">, 33(3):41, 2012.</span></div><div class="" style="margin: 0cm 0cm 0.0001pt; text-align: justify;"><span lang="EN-US" class=""><font face="Times" class="">- L. G. Valiant. Three Problems in Computer Science. <i class="">Journal of the ACM</i>, 50(1):96-99,2003. <o:p class=""></o:p></font></span></div><div class="" style="margin: 0cm 0cm 0.0001pt; text-align: justify;"><span lang="EN-US" class="" style="font-family: Times;"> </span></div><h3 class=""><span lang="EN-US" class="">Candidate profile</span></h3><p class="MsoNormal" style="text-align: justify;"><span lang="EN-US" class="" style="font-family: Times;">The PhD is at the intersection of logic, game theory and machine learning. The ideal candidate should have a strong mathematical background and a master’s degree in Logic, </span></p><p class="MsoNormal" style="text-align: justify;"><span lang="EN-US" class="" style="font-family: Times;">Computer Science or Mathematics. Ideally, it should be familiar with propositional logic, modal logic, epistemic and temporal logics, the theory of static and sequential games as well as with basic notions of machine learning. <o:p class=""></o:p></span></p><div class="" style="text-align: justify;"><span lang="EN-US" class="" style="font-family: Times;"> </span><br class="webkit-block-placeholder"></div><h3 class=""><span lang="EN-US" class="">PhD supervisor<o:p class=""></o:p></span></h3><p class="MsoNormal"><span class="" style="text-align: justify; font-family: Times;">The PhD supervisor is Emiliano Lorini, CNRS research director at the Institut de Recherche en Informatique de Toulouse (IRIT). See </span><a href="https://www.irit.fr/~Emiliano.Lorini/" class="" style="text-align: justify;"><span class="" style="font-family: Times;">https://www.irit.fr/~Emiliano.Lorini/</span></a><span class="" style="text-align: justify; font-family: Times;"> for more information.</span></p><div class=""><br class=""></div><h3 class=""><span lang="EN-US" class="">How to apply<o:p class=""></o:p></span></h3><p class="MsoNormal" style="text-align: justify;"><span lang="EN-US" class="" style="font-family: Times;">Please email your detailed CV, a motivation letter, and transcripts of bachelor's degree and master’s degree to </span><a href="mailto:Emiliano.Lorini@irit.fr" class=""><span lang="EN-US" class="" style="font-family: Times;">Emiliano.Lorini@irit.fr</span></a><span lang="EN-US" class="" style="font-family: Times;">. </span></p><div class=""><br class=""></div><div class=""><span class="Titre3Car" style="text-align: justify;"><span lang="EN-US" class=""><b class="">APPLICATION DEADLINE FOR FULL CONSIDERATION</b></span></span><span lang="EN-US" class="" style="text-align: justify; font-family: Times;">: <b class="">May 1st, 2023.</b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><span lang="EN-US" class="" style="text-align: justify; font-family: Times;"><b class=""><br class=""></b></span></div><div class=""><br class=""></div></body></html>