Dear all,

We will have two events next week.

(1) Monday: PhD defense and workshop

In celebration of Daira Pinto Prieto's PhD defense, there is a workshop focused on aggregation methods, uncertainty, and learning.

The workshop will begin at 9 AM at the Amsterdam University Library and will be followed by Daira's defense of her thesis, Combining Uncertain Evidence: Logic and Complexity, at 2 PM in the Agnietenkapel.

Date: Monday, November 25th 2024, 9:00-16:00

More information about the workshop and registration: https://sites.google.com/view/dpp-phd-defense-workshop More information about the defense: https://www.uva.nl/en/content/events/2024/11/combining-uncertain-evidence-lo...

(2) Thursday: LIRa session

To attend online, please use our recurring zoom link: https://uva-live.zoom.us/j/89230639823?pwd=YWJuSnJmTDhXcWhmd1ZkeG5zb0o5UT09 (Meeting ID: 892 3063 9823, Passcode: 421723)

You can find the details of the talk below.

Speaker: Fernando Raymundo Velázquez-Quesada (University of Bergen)

Date and Time: Thursday, November 28th 2024, 16:30-18:00

Venue: Online only.

Title: On distributed beliefs

Abstract. Within Epistemic Logic, the distributed knowledge of a group is typically defined as what is the case in all the worlds reachable by the intersection of the group's members' indistinguishability relations. Although this definition does not match the original intuitive idea (the logical consequences of the union of the individual member's knowledge), it still has a 'nice' behaviour. In particular, since intersection preserves the standard relational properties for knowledge (reflexivity, transitivity, Euclidicity), distributed knowledge is truthful and introspective (when the agents' individual knowledge is), and so will be the knowledge of the agents after a 'matching' communication action. Still, this strategy does not work properly when one switches to a KD45 (serial, transitive, Euclidean) notion of belief. Intersections do not preserve seriality, so the distributed beliefs of consistent agents might not be consistent, and consistent agents might become inconsistent after sharing their beliefs. The talk discusses an approach that uses maximally consistent sets of agents to deal with this issue. This is used, first, to define two forms of distributed belief that 'behave better' w.r.t inconsistencies, and then, to define a form of communication that preserves the properties of beliefs. The talk is based on ongoing work with John Lindqvist and Thomas Ågotnes.

Hope to see you there!

The LIRa team