Ryu Parish (龍)

Hello!

My name is Ryu, prounouced like kangaroo.

I am a BS/MS (Senior) student at the University of Washington, formerly (on a break) leading a NLP-SE project at PLSE advised by René Just and mentored by Sam Kaufman.

Before that, I was advised by Hannaneh Hajishirzi and mentored by Ellen Wu at the H2Lab and worked on a project about challenging language models to handle Mixed-Initiative Interactions. It was difficult but the things I learned were incredibly valuable and I am extremely thankful to have gotten to work with Ellen.

I transferred to the Paul G. Allen School of Computer Science and Engineering at the University of Washington from Tarrant County College in Fort Worth, Texas back in 2021 to pursue a BS in Computer Science.

My research interests are in Machine Learning for Code, especially in applying Large Language Models to Code Understanding. I believe that all software should be open sourced.

Here is my CV.

news

Jan 15, 2024	Returning (again) to AWS as a Software Engineer Intern for Summer 2024!
Nov 1, 2022	Joined HuggingFace’s BigCode project
Oct 6, 2022	InSCIt is conditionally accepted into TACL!
Sep 18, 2022	Returning to AWS as a Software Engineer Intern for Summer 2023!
Jul 13, 2022	Joined the BS/MS program at the Allen School!
Jul 2, 2022	Finished my first nlp project in collaboration with Ellen Wu. Ellen is the best.
Mar 16, 2022	Police called on Black CS student, Hawa Drammeh
Nov 5, 2021	Began my first research project under Ellen Wu @ H2Lab.
Sep 25, 2021	Amazon SWE internship offer accepted!
Sep 16, 2021	Joined Google Research CSRMP Class of 2021 for Machine Perception. (now a Class B alumni)

selected publications

MutantSelection

Leveraging Language Models to Prioritize Mutants

In Progress 2022
InSCIt

INSCIT: Information-Seeking Conversations with Mixed-Initiative Interactions

Transactions of the ACL (conditional) 2022

Abs HTML

In an information-seeking conversation, a user converses with an agent to ask a series of ques- tions that can often be under- or over-specified. An ideal agent would first identify that they were in such a situation by searching through their underlying knowledge source and then ap- propriately interacting with a user to resolve it. However, most existing studies either fail to or artificially incorporate such agent-side initiatives. In this work, we present INSCIT (pronounced Insight), a dataset for Information- Seeking Conversations with mixed-initiative Interactions. It contains a total of 4.7K user- agent turns from 805 human-human conversa- tions where the agent searches over Wikipedia and either asks for clarification or provides rel- evant information to address user queries. We define two subtasks, namely evidence passage identification and response generation, as well as a new human evaluation protocol to assess the model performance. We report results of two strong baselines based on state-of-the-art models of conversational knowledge identifi- cation and open-domain question answering. Both models significantly underperform hu- mans and fail to generate coherent and infor- mative responses, suggesting ample room for improvement in future studies.