ILLC-LIRa May 2021

illc-lira@list.uva.nl

1 participants
8 discussions

next week: DIEP talk (Velimir Ilić) and regular LIRa session (Guillermo Menéndez Turata)
by LIRa Seminar 28 May '21

28 May '21

Dear all, Next Thursday, June 3rd, we have two events of interest to the LIRa audience. The first is the DIEP talk of Velimir Ilić, from 11:00 to 12:30. The second is our regular LIRa session with Guillermo Menéndez Turata, from 16:30 to 18:00. You can find the details for both events, as well as their zoom links, below. ----------------------------------------------------- Date and time: Thursday, June 3rd 2021, 11:00-12:30, Amsterdam time. Speaker: Velimir Ilić (Mathematical Institute of the Serbian Academy of Sciences and Arts) Title: An overview and characterization of generalized information measures Abstract: The aim of this talk is to present a comprehensive classification of the main entropic forms introduced in the last fifty years within statistical physics and information theory and to review the fundamental questions about the meaning of information. I will particularly focus on axiomatic approaches to the characterization of various generalizations of the Shannon entropy, such as the Rényi, the Tsallis, the Sharma-Mittal, and the Sharma-Mittal-Taneja entropies, as well as the more general classes of pseudo-additive entropies with a well-defined mathematical and information theoretic structure. Finally, I will point out possible applications of these measures in communication theory, statistical inference and complex systems modeling. Zoom link: https://uva-live.zoom.us/j/85608909905 ----------------------------------------------------- Speaker: Guillermo Menéndez Turata Date and Time: Thursday, June 3rd 2021, 16:30-18:00, Amsterdam time. Venue: online. Title: Uniform interpolation from cyclic proofs: the case of modal mu-calculus. Abstract. Uniform interpolation was first established for the modal mu-calculus by D’Agostino and Hollenberg via a combination of semantic and syntactic methods. A natural question is whether a purely syntactical proof of this result, in the style of Pitt’s seminal work on uniform interpolation for intuitionistic logic, can be produced. One possible reason Pitt’s method has not been applied to the modal mu-calculus is that it requires a setting where uniform interpolation can be proved by induction on cut-free derivations in a finitary system. While it is feasible to design cut-free calculi for fixpoint modal logics, they are often infinitary, i.e., derivations may have infinite branches. In this talk we apply Pitt’s technique to the modal mu-calculus and show how to construct uniform interpolants from cyclic derivations in an annotated goal-oriented proof system introduced by Jungteerapanich and Stirling. Recurring zoom link: https://uva-live.zoom.us/j/92907704256?pwd=anY3WkFmQVhLZGhjT2JXMlhjQVl1dz09 (Meeting ID: 929 0770 4256, Passcode: 036024) Hope to see you there! The LIRa team

1 0

tomorrow: DIEP talk (Olivier Roy) and LIRa session (Francesca Zaffora Blando)
by LIRa Seminar 26 May '21

26 May '21

Dear all, Tomorrow, May 27th, we have two events of interest to the LIRa audience. The first is the DIEP talk of Olivier Roy, from 11:00 to 12:30. The second is our regular LIRa session with Francesca Zaffora Blando, from 16:30 to 18:00. You can find the details for both events, as well as their zoom links, below. ----------------------------------------------------------------- Date and time: Thursday, May 27th 2021, 11:00-12:30, Amsterdam time. Speaker: Olivier Roy (Bayreuth University) Title: Deliberation, Coherent Aggregation, and Anchoring Abstract: In this talk we will present a number of results stemming from a computational model of collective attitude formation through a combination of group deliberation and aggregation. In this model the participants repeatedly exchange and update their preferences over small sets of alternatives, until they reach a stable preference profile. When they do so the collective attitude is computed by pairwise majority voting. The model shows, on the one hand, that rational preference change can fill an existing gap in known mechanisms purported to explain how deliberation can help avoiding incoherent group preferences. On the other hand, the model also reveals that when the participants are sufficiently biased towards their own opinion, deliberation can actually create incoherent group rankings, against the received view. The model suggests furthermore that rational deliberation can exhibit high levels of path dependencies or \\\"anchoring\\\", where the group opinion is strongly dependent on the order in which the participants contribute to the discussion. We will finish by discussing possible trade-offs between such positive and negative features of group deliberation. Zoom link: https://uva-live.zoom.us/j/85608909905 ----------------------------------------------------------------- Speaker: Francesca Zaffora Blando Date and Time: Thursday, May 27th 2021, 16:30-18:00, Amsterdam time. Venue: online. Title: Weak merging of opinions for computationally limited agents. Abstract. A standard objection to subjective Bayesianism is that appealing to subjective probabilities threatens the objectivity of scientific inquiry. A standard Bayesian response to this charge relies on merging-of-opinions theorems: a family of results which establish that, as long as their respective priors are sufficiently compatible, two Bayesian agents with differing initial beliefs are guaranteed to almost surely reach a consensus with increasing evidence. So, objectivity can be recovered in the form of intersubjective agreement. One of the most well-known such results is the Blackwell-Dubins Theorem, which shows that Bayesian conditioning leads to a strong form of merging of opinions, provided that the agents agree on probability zero events to begin with?i.e., provided that their priors are mutually absolutely continuous. Since absolute continuity is a rather strong form of compatibility between priors, it is natural to wonder whether merging of opinions?and what type of merging of opinions?can be achieved with weaker assumptions. In this talk, I will address this question from the perspective of computationally limited Bayesian agents: agents whose priors are computable. I will argue that, for computable Bayesian learners, it is natural to appeal to the theory of algorithmic randomness?a branch of computability theory aimed at characterizing the concept of effective measure-theoretic typicality?to define notions of compatibility between priors. We will see that the proposed notions of compatibility induced by algorithmic randomness naturally correspond to restricted forms of absolute continuity. Then, I will show that some of these notions, while too weak to ensure merging of opinions in the strong sense of Blackwell and Dubins, nonetheless suffice to attain a weaker type of merging, first studied by Kalai and Lehrer, which only requires reaching a consensus over finite-horizon events. Recurring zoom link: https://uva-live.zoom.us/j/92907704256?pwd=anY3WkFmQVhLZGhjT2JXMlhjQVl1dz09 (Meeting ID: 929 0770 4256, Passcode: 036024) Hope to see you there! The LIRa team

1 0

LIRa-related VvL Logic at Large Lectures: And Logic Begat Computer Science
by LIRa Seminar 24 May '21

24 May '21

Dear all, This is to remind you that apart from the events scheduled for this Thursday (DIEP talk by Olivier Roy and regular LIRa seminar by Francesca Zaffora Blando), there will also be a LIRa-related event on Friday, namely the VvL Logic at Large Lecture: And Logic Begat Computer Science. You can find the details below. *********************************************************************** * VvL Logic at Large Lectures: And Logic Begat Computer Science * * When: Friday 28 May 2021, 15:30 - 17:45 / 18:45 * * Where: Online (Zoom, Gather.town) * *********************************************************************** To mark its relaunch the VvL (Dutch Association for Logic and Philosophy of Exact Sciences) has the privilege to announce its first outreach event in 2021. It will take place on *Friday 28 May 2021*. We are very pleased to announce that Professor Moshe Y. Vardi (Rice University) will give a public lecture entitled \\\"And Logic Begat Computer Science\\\" (see the abstract below). The event will take place online using Zoom. The talk will be followed by a short session where three commentators from different areas of logic and computer science will react to the talk and kick off the discussion which will then transition into a general Q&A with the audience. We are delighted to announce that the commentators will be: - Johan van Benthem (UvA, Stanford, Tsinghua), - Frank van Harmelen (VU Amsterdam, Wuhan), - Marieke Huisman (U Twente). The outreach event will be concluded by a social gathering on the virtual platform Gather.Town. Registration is free, but necessary to receive links to Zoom and Gather.Town. For registration and more information, please visit: http://www.verenigingvoorlogica.nl/Activiteiten/ Programme: 15:30-15:50 Gathering/Informal discussion 15:50-16:00 An update on VvL and its activities 16:00-17:00 Public Lecture by Moshe Y. Vardi (Rice University): And Logic Begat Computer Science 17:00-17:45 Questions and discussion with commentators 17:45-18:45 Social Event on Gather.Town For any questions, please feel free to contact the organisers: - Nick Bezhanishvili, University of Amsterdam (N.Bezhanishvili(a)uva.nl) - Davide Grossi, University of Groningen (D.Grossi(a)rug.nl) - Helle Hvid Hansen, University of Groningen (H.H.Hansen(a)rug.nl) - Dominik Klein, Utrecht University (D.Klein(a)uu.nl) - Peter van Ormondt, University of Amsterdam (P.vanOrmondt(a)uva.nl) ------------------------------------------------------------------------- And Logic Begat Computer Science Moshe Y. Vardi Rice University Abstract: During the past fifty years there has been extensive, continuous, and growing interaction between logic and computer science. In fact, logic has been called \\\"the calculus of computer science\\\". The argument is that logic plays a fundamental role in computer science, similar to that played by calculus in the physical sciences and traditional engineering disciplines. Indeed, logic plays an important role in areas of computer science as disparate as architecture (logic gates), software engineering (specification and verification), programming languages (semantics, logic programming), databases (relational algebra and SQL), artificial intelligence (automated theorem proving), algorithms (complexity and expressiveness), and theory of computation (general notions of computability). This non-technical talk will provide an overview of the unusual effectiveness of logic in computer science by surveying the history of logic in computer science, going back all the way to Aristotle and Euclid, and showing how logic actually gave rise to computer science. *********************************************************************** Hope to see you there! The LIRa team

1 0

LIRa Session: Francesca Zaffora Blando
by LIRa Seminar 21 May '21

21 May '21

Dear all, We will have our next LIRa session on Thursday, May 27th. Our speaker is Francesca Zaffora Blando. You can find the details of the talk below. We will use our recurring zoom link: https://uva-live.zoom.us/j/92907704256?pwd=anY3WkFmQVhLZGhjT2JXMlhjQVl1dz09 (Meeting ID: 929 0770 4256, Passcode: 036024) Speaker: Francesca Zaffora Blando Date and Time: Thursday, May 27th 2021, 16:30-18:00, Amsterdam time. Venue: online. Title: Weak merging of opinions for computationally limited agents. Abstract. A standard objection to subjective Bayesianism is that appealing to subjective probabilities threatens the objectivity of scientific inquiry. A standard Bayesian response to this charge relies on merging-of-opinions theorems: a family of results which establish that, as long as their respective priors are sufficiently compatible, two Bayesian agents with differing initial beliefs are guaranteed to almost surely reach a consensus with increasing evidence. So, objectivity can be recovered in the form of intersubjective agreement. One of the most well-known such results is the Blackwell-Dubins Theorem, which shows that Bayesian conditioning leads to a strong form of merging of opinions, provided that the agents agree on probability zero events to begin with—i.e., provided that their priors are mutually absolutely continuous. Since absolute continuity is a rather strong form of compatibility between priors, it is natural to wonder whether merging of opinions—and what type of merging of opinions—can be achieved with weaker assumptions. In this talk, I will address this question from the perspective of computationally limited Bayesian agents: agents whose priors are computable. I will argue that, for computable Bayesian learners, it is natural to appeal to the theory of algorithmic randomness—a branch of computability theory aimed at characterizing the concept of effective measure-theoretic typicality—to define notions of compatibility between priors. We will see that the proposed notions of compatibility induced by algorithmic randomness naturally correspond to restricted forms of absolute continuity. Then, I will show that some of these notions, while too weak to ensure merging of opinions in the strong sense of Blackwell and Dubins, nonetheless suffice to attain a weaker type of merging, first studied by Kalai and Lehrer, which only requires reaching a consensus over finite-horizon events. Hope to see you there! The LIRa team

1 0

LIRa-related events: DIEP talks by Sven Banisch (tomorrow) and Olivier Roy (May 27)
by LIRa Seminar 19 May '21

19 May '21

Dear all, Apart from the regular LIRa session scheduled for tomorrow (Nina Gierasimczuk) and advertised earlier today, we would like to bring to your attention the following two talks of DIEP (Dutch Institute for Emergent Phenomena) which might be of interest to the LIRa audience. The talks are by Sven Banisch (tomorrow) and by Olivier Roy (May 27). You can find the details of the talks, as well as the zoom links, below. Date and time: Thursday, May 20th 2021, 11:00-12:30, Amsterdam time. Speaker: Sven Banisch (Max Planck Institute, Leipzig) Title: Social Feedback Theory: Modeling collective opinion phenomena by learning from the feedback of others Abstract: Humans are sensitive to social approval and disapproval. The feedback that others provide on our expressions of opinion is an important driver for adaptation and change. Social feedback theory provides a framework for modeling collective opinion processes based on these principles. The theory departs from previous models by differentiating an externally expressed opinion from an internal evaluation of it. Opinion dynamics is conceived as repeated games that agents play within their social network and to which they adapt by reward-based learning. Within this setting, game theoretic notions of equilibrium can be used to characterize structural conditions for qualitatively different regimes of collective opinion expression. In this talk, I aim for a broad perspective and discuss two models addressing emergent phenomena such as polarization and collective silence. Zoom link: https://uva-live.zoom.us/j/85608909905 ------------------------------------------------------------------- Date and time: Thursday, May 27th 2021, 11:00-12:30, Amsterdam time. Speaker: Olivier Roy (Bayreuth University) Title: Deliberation, Coherent Aggregation, and Anchoring Abstract: In this talk we will present a number of results stemming from a computational model of collective attitude formation through a combination of group deliberation and aggregation. In this model the participants repeatedly exchange and update their preferences over small sets of alternatives, until they reach a stable preference profile. When they do so the collective attitude is computed by pairwise majority voting. The model shows, on the one hand, that rational preference change can fill an existing gap in known mechanisms purported to explain how deliberation can help avoiding incoherent group preferences. On the other hand, the model also reveals that when the participants are sufficiently biased towards their own opinion, deliberation can actually create incoherent group rankings, against the received view. The model suggests furthermore that rational deliberation can exhibit high levels of path dependencies or “anchoring”, where the group opinion is strongly dependent on the order in which the participants contribute to the discussion. We will finish by discussing possible trade-offs between such positive and negative features of group deliberation. Zoom link: https://uva-live.zoom.us/j/85608909905 Hope to see you there! The LIRa team

1 0

LIRa Session: Nina Gierasimczuk
by LIRa Seminar 19 May '21

19 May '21

Dear all, We will have our next LIRa session tomorrow. We will use our recurring zoom link: https://uva-live.zoom.us/j/92907704256?pwd=anY3WkFmQVhLZGhjT2JXMlhjQVl1dz09 (Meeting ID: 929 0770 4256, Passcode: 036024). You can find the details of the talk below. Speaker: Nina Gierasimczuk Date and Time: Thursday, May 20th 2021, 16:30-18:00, Amsterdam time. Venue: online. Title: Multi-agent language coordination and cognitive semantics of quantity terms Abstract. Natural languages vary in their quantity expressions, but the variation seems to be constrained by general properties, so-called universals. Explanations thereof have been sought among constraints of human cognition, communication, complexity, and pragmatics. In this work, we examine whether the perceptual constraints of approximate number sense (ANS) contribute to the development of two universals in the semantic domain of quantities: monotonicity and convexity. Using a state-of-the-art multi-agent language coordination model (originally applied to colour terms) we evolve communicatively usable quantity terminologies. We compare the degrees of convexity and monotonicity of languages evolvingin populations of agents with and without ANS. The results suggest that ANS supports the development of monotonicity and, to a lesser extent, convexity. This is a joint work with Dariusz Kalociński (University of Warsaw), Franciszek Rakowski (Samsung Research & Development Institute Poland) and Jakub Uszyński (National Information Processing Institute, Poland). It was generously supported by N.G.’s National Science Centre Poland grant no. 2015/19/B/HS1/03292. Hope to see you there! The LIRa team

1 0

LIRa Session: Nina Gierasimczuk
by LIRa Seminar 14 May '21

14 May '21

Dear all, We will have our next LIRa session on Thursday, May 20th. Our speaker is Nina Gierasimczuk. You can find the details of the talk below. We will use our recurring zoom link: https://uva-live.zoom.us/j/92907704256?pwd=anY3WkFmQVhLZGhjT2JXMlhjQVl1dz09 (Meeting ID: 929 0770 4256, Passcode: 036024) Speaker: Nina Gierasimczuk Date and Time: Thursday, May 20th 2021, 16:30-18:00, Amsterdam time. Venue: online. Title: Multi-agent language coordination and cognitive semantics of quantity terms Abstract. Natural languages vary in their quantity expressions, but the variation seems to be constrained by general properties, so-called universals. Explanations thereof have been sought among constraints of human cognition, communication, complexity, and pragmatics. In this work, we examine whether the perceptual constraints of approximate number sense (ANS) contribute to the development of two universals in the semantic domain of quantities: monotonicity and convexity. Using a state-of-the-art multi-agent language coordination model (originally applied to colour terms) we evolve communicatively usable quantity terminologies. We compare the degrees of convexity and monotonicity of languages evolvingin populations of agents with and without ANS. The results suggest that ANS supports the development of monotonicity and, to a lesser extent, convexity. This is a joint work with Dariusz Kalociński (University of Warsaw), Franciszek Rakowski (Samsung Research & Development Institute Poland) and Jakub Uszyński (National Information Processing Institute, Poland). It was generously supported by N.G.’s National Science Centre Poland grant no. 2015/19/B/HS1/03292. Hope to see you there! The LIRa team

1 0

tomorrow: LIRa-related PEPTalk #6 and LIRa session with Hein Duijf
by LIRa Seminar 05 May '21

05 May '21

Dear all, This is to remind you that the following two events will take place tomorrow: First, the PEPTalk on Responsible AI, Machine Ethics, and Logic, featuring Jan Broersen and Marija Slavkovik, and moderated by Aybüke Özgün. It will take place at 12:00 -13:00. The details of the event and the registration can be found here: https://www.uva.nl/en/shared-content/faculteiten/en/faculteit-der-geesteswe… Second, our regular LIRa session. We will use our recurring zoom link: https://uva-live.zoom.us/j/92907704256?pwd=anY3WkFmQVhLZGhjT2JXMlhjQVl1dz09 (Meeting ID: 929 0770 4256, Passcode: 036024). You can find the details of the LIRa talk below. Speaker: Hein Duijf Date and Time: Thursday, May 6th 2021, 16:30-18:00, Amsterdam time. Venue: online. Title: Should one be open-minded? Abstract. I investigate the common intuition that open-mindedness is virtuous, while its opposite – close-mindedness – is vicious. I defend a positive claim that open-mindedness sometimes helps us achieve some basic epistemic goals. However, I also defend a negative claim that open-mindedness sometimes hinders the achievement of some basic epistemic goals. Open-mindedness is only epistemically beneficial if one has decent evaluative skills. These evaluative skills can be understood in several ways: as the social skill to determine whether another person is trustworthy or has mutual interests; or as the evaluative skill to determine whether certain argument or piece of evidence is truth-conducive. In the absence of sufficient evaluative skills and for low degrees of open mindedness, it seems plausible that open-mindedness will not be epistemically fruitful. Hope to see you there! The LIRa team

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

ILLC-LIRa May 2021