[4eyes] FW: ACL Practice Talk on Inverse Reinforcement Learning

[Matthew Turk]

FYI
 
From: William Wang [mailto:william at cs.ucsb.edu] 
Sent: Monday, June 11, 2018 10:40 AM
To: nlp at lists.cs.ucsb.edu; ucsb-ml at lists.cs.ucsb.edu
Cc: Matthew Turk <mturk at cs.ucsb.edu>; B. S. Manjunath <manj at ece.ucsb.edu>; Miguel Eckstein <miguel.eckstein at psych.ucsb.edu>; Xin Wang <xwang at cs.ucsb.edu>
Subject: ACL Practice Talk on Inverse Reinforcement Learning
 
Hi all,
 
I'd like to invite you to a practice talk this Thursday.
 
Time: Thu 06/14 11am.
Location: HFH 1152.
Speaker: Xin Wang.
 
Title: No Metrics Are Perfect: Adversarial Reward Learning for Visual Storytelling
 
Abstract: Though impressive results have been achieved in visual captioning, the task of generating abstract stories from photo streams is still a little-tapped problem. Different from captions, stories have more expressive language styles and contain many imaginary concepts that do not appear in the images. Thus it poses challenges to behavioral cloning algorithms. Furthermore, due to the limitations of automatic metrics on evaluating story quality, reinforcement learning methods with hand-crafted rewards also face difficulties in gaining an overall performance boost. Therefore, we propose an Adversarial REward Learning (AREL) framework to learn an implicit reward function from human demonstrations, and then optimize policy search with the learned reward function. Though automatic evaluation indicates slight performance boost over state-of-the-art (SOTA) methods in cloning expert behaviors, human evaluation shows that our approach achieves significant improvement in generating more human-like stories than SOTA systems.
 
All are welcome!
 
William

 
-- 
William Wang
 
Assistant Professor
Department of Computer Science
University of California, Santa Barbara
https://www.cs.ucsb.edu/~william
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cs.ucsb.edu/pipermail/ilab-users/attachments/20180611/f5c0df03/attachment-0001.html>