Dear all,
This Thursday we have two LIRa-related events.
One is the announced LIRa talk by Andres Occhipinti, with title: Learning to Act and Observe in Partially Observable Domains. This will take place at the regular time (16:30-18:00), and you can use our recurring zoom link: https://uva-live.zoom.us/j/92907704256?pwd=anY3WkFmQVhLZGhjT2JXMlhjQVl1dz09 (Meeting ID: 929 0770 4256, Passcode: 036024). You can find more details below.
The second event (taking place in fact earlier on Thursday) is the workshop of DIEP (Dutch Institute for Emergent Phenomena), between 13:00-16:00, which includes a talk by ILLC-based DIEP fellow Soroush Rafiee Rad, who is soon to join our group, and who will give a talk with title: Characterizing Probabilistic Models with a Symmetry Axiom" (at 14:20). The zoom link for this second event is: https://uva-live.zoom.us/j/87881838105 More details follow below.
----------------------------------------------------------------------------------------- Regular LIRa TALK: Andres Occipinti Liberman
Date and Time: Thursday, November 26th 2020, 16:30-18:00, Amsterdam time.
Title: Learning to Act and Observe in Partially Observable Domains
Abstract: We consider a learning agent in a partially observable domain (or environment), with which the agent has never interacted before. The agent wishes to learn a representation of the domain dynamics: how the agent\'s actions affect the state of the domain and what becomes observable as a result of such actions. To produce such knowledge, the learner has access to experience gathered by taking actions in the domain and observing their results.
We present algorithms for learning \"as much as possible\" (in a well-defined sense) both about what is directly observable and about what actions do in the domain, given the learner\'s observational constraints. We differentiate the level of domain knowledge attained by each algorithm, and characterize the type of observations required to reach it. The algorithms use dynamic epistemic logic (DEL) to represent the learned domain information symbolically. Our work continues that of Bolander and Gierasimczuk (2015), which developed DEL-based learning algorithms based to learn domain information in fully observable domains.
This is joint work with Thomas Bolander and Nina Gierasimczuk. ---------------------------------------------------------------------------------------
LIRa-related event: DIEP Workshop (including talk by Soroush Rafiee Rad)
Date and Time: Thursday, November 26th 2020, 13:30-16:00, Amsterdam time. Soroush's talk is at 14:20.
On Thursday, the 26th of November from 1pm to 4pm, the Dutch Institute for Emergent Phenomena (DIEP) is organising a series of virtual talks by the DIEP fellows and researchers who will be joining DIEP@UvA soon. ILLC is the host institute for one of these fellows, Soroush Rafiee Rad, who will give a talk on ``Characterizing Probabilistic Models with a Symmetry Axiom". The topics covered by the talks include multiscale modelling of reaction-diffusion networks, self-learning algorithms in duopoly, axiomatic approaches to probabilistic models, and many-body stochastic systems. Everybody is welcome to attend, and the Zoom link is https://uva-live.zoom.us/j/87881838105 You will have the opportunity to meet DIEP researchers and hear more about upcoming activities. More information (including full programme and abstracts) is available here: https://www.d-iep.org/diepuvakickoff
Hope to see you there!
The LIRa team