nl-uiuc AT lists.cs.illinois.edu
Subject: Natural language research announcements
List archive
- From: "Alexandre Klementiev" <klementi AT uiuc.edu>
- To: nl-uiuc AT cs.uiuc.edu, aivr AT cs.uiuc.edu, dais AT cs.uiuc.edu, cogcomp AT cs.uiuc.edu
- Subject: [nl-uiuc] Talk tomorrow (Piotr J. Gmytrasiewicz).
- Date: Wed, 5 Dec 2007 14:19:27 -0600
- List-archive: <http://lists.cs.uiuc.edu/pipermail/nl-uiuc>
- List-id: Natural language research announcements <nl-uiuc.cs.uiuc.edu>
Forwarding a talk announcement from Kiran Lakkaraju.
Alex.
Alex, can you send this out to the AI groups? Prof. Piotr
Gmytrasiewicz is giving a talk on POMDPs that will be of interest to
many in the department. The talk is tomorrow though, so quick
advertisement is necessary!
-Kiran
------------------------------------------------------
Title: Interactive Partially Observable Markov Decision Processes
Prof. Piotr J. Gmytrasiewicz
University of Illinois - Chicago
3pm, December 6th, 2007 in SC 3403
Abstract:
The talk will extend the framework of partially observable Markov
decision processes (POMDPs) to multi-agent settings by incorporating
the notion of agent models into the state space. Agents maintain
beliefs over physical states of the environment and over models of
other agents, and they use Bayesian updates to maintain their beliefs
over time. The solutions map belief states to actions. Models of other
agents may include their belief states and are related to agent types
considered in games of incomplete information. We express the agents'
autonomy by postulating that their models are not directly manipulable.
IPOMDPs complement a traditional approach to interactive settings
which uses Nash
equilibria as a solution paradigm. We seek to avoid some of the
drawbacks of equilibria which may be non-unique and do not capture
off-equilibrium behaviors. We do so at the cost of having to
represent, process and continuously revise models of other agents.
Since the agent's beliefs may be arbitrarily nested, the optimal
solutions to decision making problems are only asymptotically
computable.
Gmytrasiewicz is giving a talk on POMDPs that will be of interest to
many in the department. The talk is tomorrow though, so quick
advertisement is necessary!
-Kiran
------------------------------------------------------
Title: Interactive Partially Observable Markov Decision Processes
Prof. Piotr J. Gmytrasiewicz
University of Illinois - Chicago
3pm, December 6th, 2007 in SC 3403
Abstract:
The talk will extend the framework of partially observable Markov
decision processes (POMDPs) to multi-agent settings by incorporating
the notion of agent models into the state space. Agents maintain
beliefs over physical states of the environment and over models of
other agents, and they use Bayesian updates to maintain their beliefs
over time. The solutions map belief states to actions. Models of other
agents may include their belief states and are related to agent types
considered in games of incomplete information. We express the agents'
autonomy by postulating that their models are not directly manipulable.
IPOMDPs complement a traditional approach to interactive settings
which uses Nash
equilibria as a solution paradigm. We seek to avoid some of the
drawbacks of equilibria which may be non-unique and do not capture
off-equilibrium behaviors. We do so at the cost of having to
represent, process and continuously revise models of other agents.
Since the agent's beliefs may be arbitrarily nested, the optimal
solutions to decision making problems are only asymptotically
computable.
- [nl-uiuc] Talk tomorrow (Piotr J. Gmytrasiewicz)., Alexandre Klementiev, 12/05/2007
Archive powered by MHonArc 2.6.16.