[Colloquium] Fwd: NLP Seminar 3/22 Friday 11:00 am - Zhewei Sun

[Jiawei Zhou]

Forwarding the following NLP talk information at TTIC on Friday morning at
11 am. The speaker, Zhewei Sun, will join TTIC as a research assistant
professor in the fall. The talk will be in-person at TTIC 529 and also on
zoom
<https://uchicago.zoom.us/j/98297764499?pwd=ajNQSTZnMHRmMENkd1hjdjlNeW1xdz09>,
and you can sign up to meet with the speaker here
<https://docs.google.com/spreadsheets/d/1GLDLCBdsVbwSDeSyT6JctJq0JyrWclfFgkoHz69V4hI/edit?usp=sharing>
!


Best,

Jiawei (Joe) Zhou
TTIC
jzhou at ttic.edu


---------- Forwarded message ---------
From: Jiawei Zhou <jzhou at ttic.edu>
Date: Mon, Mar 18, 2024 at 12:04 PM
Subject: NLP Seminar 3/22 Friday 11:00 am - Zhewei Sun
To: <nlp at lists.uchicago.edu>


Hi all,

Welcome back from the spring break! We will resume our NLP seminar on
Fridays 11:00 am - 12:00 pm at TTIC Room 529 (and on zoom
<https://uchicago.zoom.us/j/98297764499?pwd=ajNQSTZnMHRmMENkd1hjdjlNeW1xdz09>).
For this week, we are excited to have Zhewei Sun from the University of
Toronto visiting us in-person on 3/22. Zhewei will join TTIC this fall as a
research assistant professor! If you want to meet him you can sign up here
<https://docs.google.com/spreadsheets/d/1GLDLCBdsVbwSDeSyT6JctJq0JyrWclfFgkoHz69V4hI/edit?usp=sharing>.
Below is the talk information:


Title:

Contextualizing Natural Language Agents: The Case of Slang



Abstract:

Large language models (LLMs) have recently emerged as an effective tool for
enhancing human productivity. Although conversational artificial
intelligence (AI) agents such as ChatGPT produces natural language that is
much more human-like compared to its predecessors, they still lack the
ability to situate themselves in a communicative context like humans do.
For instance, humans may opt to use informal language such as slang in a
conversation to express a close bond with the other party. In this talk, I
will illustrate how natural language processing (NLP) techniques can enable
automatic processing of slang. The first part of my talk will cover
knowledge-driven approaches that inject linguistic and cognitive knowledge
about slang into foundational NLP models for (1) generation, (2)
interpretation, and (3) modeling slang semantic variation. Next, I will
discuss emerging data-driven approaches based on large language models,
examining the extent of knowledge LLMs have acquired about slang and how
such knowledge may have been obtained. Finally, I will discuss potential
future directions to further enhance an LLM’s ability to process
contextualized language.


Bio:

Zhewei Sun is a Ph.D. Candidate in the Department of Computer Science at
the University of Toronto, working with Professor Yang Xu. His research is
focused on natural language processing (NLP) of informal language, with a
particular interest in the automatic processing of slang. His research
explores both data-driven NLP approaches based on deep learning, as well as
knowledge-driven approaches that combine traditional NLP techniques with
linguistic and cognitive knowledge about human language. His work has been
recognized at top-tier NLP venues, including EMNLP, NAACL, and TACL.
Zhewei’s work has been supported in part by Amazon Alexa AI and the Queen
Elizabeth II Graduate Scholarship in Science and Technology. Prior to
joining the University of Toronto, Zhewei received an M.Sc. in Computer
Science from Georgia Institute of Technology and a B.Sc. in Computer
Science from the University of Waterloo.




Best,

Jiawei (Joe) Zhou
TTIC
jzhou at ttic.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20240318/e088808a/attachment.html>