BEGIN:VCALENDAR
METHOD:PUBLISH
PRODID:-//Apple Computer\, Inc//iCal 1.0//EN
X-WR-CALNAME;VALUE=TEXT:USC
VERSION:2.0
BEGIN:VEVENT
DESCRIPTION:Title: Understanding Goal-oriented Reinforcement Learning\n
\n
Defense committee: Haipeng Luo, David Kempe, Ashutosh Nayyar, Rahul Jain\n
\n
Date: January 12, 2022. 1 p.m. - 2 p.m. PST\n
\n
Zoom meeting link: https://usc.zoom.us/j/93602803008\n
\n
Abstract: \n
Reinforcement Learning (RL) studies how an agent learns to behave optimally in an unknown environment. One challenge in applying RL in practice is task specification, that is, how to inform the algorithm of the task we want it to solve. My research focuses on solving the problem of task specification in goal-oriented reinforcement learning (GoRL), whose objective is to reach a goal state with the smallest possible cost. Unlike standard RL that focuses solely on cost minimization, GoRL has dual objectives: 1) reach the goal state and 2) minimize the cost. \n
\n
Standard RL models such as the finite-horizon model or the discounted model often have difficulty in specifying tasks for GoRL, which leads to heavy engineering efforts in practice.\n
\n
To tackle this issue, we study learning in a Markov Decision Process named stochastic shortest path (SSP), which exactly captures the dual objectives of GoRL. We focus on developing practical learning algorithms for SSP. Specifically, we study the PAC learning setting for SSP, and develop various reduction schemes that connect SSP to the simpler finite-horizon model. Our reduction schemes help to develop optimal and efficient online learning algorithms for SSP.
SEQUENCE:5
DTSTART:20230112T130000
LOCATION:
DTSTAMP:20230112T130000
SUMMARY:PhD Defense - Liyu Chen
UID:EC9439B1-FF65-11D6-9973-003065F99D04
DTEND:20230112T140000
END:VEVENT
END:VCALENDAR