<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
</head>
<body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
This is an announcement of Chao-Chun Hsu's Candidacy Exam.<br class="">
===============================================<br class="">
Candidate: Chao-Chun Hsu<br class="">
<br class="">
Date: Thursday, December 07, 2023<br class="">
<br class="">
Time: 1 pm CST<br class="">
<br class="">
Remote Location: <a href="https://uchicago.zoom.us/j/7554197922?pwd=UXpHQXJ5ZDFJWExmSldMdVZDRHpyQT09" class="">https://uchicago.zoom.us/j/7554197922?pwd=UXpHQXJ5ZDFJWExmSldMdVZDRHpyQT09</a><br class="">
<br class="">
Location: JCL 390<br class="">
<br class="">
Title: Understand and Improving Human Decision Making Through Texts<br class="">
<br class="">
Abstract: Reading human-written texts allows us to understand complex ideas and experiences. For example, physician notes tell us about the patient's history, and restaurant reviews reflect the customers' experience. However, given the large volume of text
on certain topics, it becomes impossible to thoroughly evaluate all documents and construct a comprehensive understanding. This challenge opens up an interesting question: Can we construct useful summaries to support prediction and decision-making by selectively
focusing on the most informative content?<br class="">
<br class="">
In the first part of this proposal, I use clinical outcome prediction tasks using clinical notes from MIMIC-III to demonstrate that not all notes are necessary for accurate outcome prediction by characterizing the value of information contained in the clinical
notes.<br class="">
<br class="">
In the second part, I propose a novel extractive summarization method called decision-focused summarization. This method aims to extract the most relevant information from source documents to support human decision-making, using future rating prediction from
Yelp restaurant reviews as a testbed.<br class="">
<br class="">
Advisors: Chenhao Tan<br class="">
<br class="">
Committee Members: Chenhao Tan, Mina Lee, Ziad Obermeyer<br class="">
<div class=""><br class="Apple-interchange-newline">
</div>
<br class="">
</body>
</html>