Zixian Ma

Ph.D. Student @ University of Washington

zixianma [at] cs.washington.edu

Hi! I am a first-year PhD student at the University of Washington in Computer Science and Engineering supervised by Prof. Ranjay Krishna. Between Aug 2022 and May 2023, I interned at Google Research and worked at Meta. Before that, I graduated from Stanford University with B.S. (Honors) and M.S. degrees in Computer Science, where I did research in the Stanford Vision and Learning Lab and HCI group under the supervision of the wonderful Prof. Fei-Fei Li and Prof. Michael Bernstein.

Recent highlights

[Sep, 2023] Our paper SugarCrepe: Fixing Hackable Benchmarks for Vision-Language Compositionality has been accepted to NeurIPS 2023 (Datasets & Benchmarks Track)!
[Mar, 2023] Our paper CREPE: Can Vision-Language Foundation Models Reason Compositionally? has been accepted to CVPR 2023 and selected as a 🌟highlight🌟!
[Jan, 2023] Our paper Model Sketching: Centering Concepts in Early-Stage Machine Learning Model Design has been accepted to CHI 2023!
[Sep, 2022] Our paper ELIGN: Expectation Alignment as a Multi-agent Intrinsic Reward has been accepted to NeurIPS 2022!
[Jun, 2022] I'm grateful that my undergraduate honors thesis Towards More Effective Multi-agent Coordination via Alignment has received the Firestone Medal for Excellence in Undergraduate Research and the Ben Wegbreit Prize for Undergraduate CS Research.

Selected publications

SugarCrepe: Fixing Hackable Benchmarks for Vision-Language Compositionality

Cheng-Yu Hsieh*, Jieyu Zhang*, Zixian Ma, Aniruddha Kembhavi, Ranjay Krishna

NeurIPS 2023 (Datasets & Benchmarks Track)

[PDF] [Code]
CREPE: Can Vision-Language Foundation Models Reason Compositionally?

Zixian Ma*, Jerry Hong*, Mustafa Omer Gul*, Mona Gandhi, Irena Gao, Ranjay Krishna

CVPR 2023 [Highlight: 10% of accepted papers / 2.5% of submissions]

[PDF]
Model Sketching: Centering Concepts in Early-Stage Machine Learning Model Design

Michelle S. Lam, Zixian Ma, Anne Li, Izequiel Freitas, Dakuo Wang, James A. Landay, Michael S. Bernstein

CHI 2023

[PDF]
ELIGN: Expection Alignment as a Multi-agent Intrinsic Reward

Zixian Ma, Rose Wang, Li Fei-Fei, Michael Bernstein, Ranjay Krishna

NeurIPS 2022

[PDF] [Code]
Openattack: An Open-source Textual Adversarial Attack Toolkit

Guoyang Zeng, Fanchao Qi, Qianrui Zhou, Tingji Zhang, Zixian Ma, Bairu Hou, Yuan Zang, Zhiyuan Liu, Maosong Sun

ACL2021 Demo

[PDF] [Code]

Projects

Sex Academy

Personalized sex education with stories.

Sex Academy is a platform that provides a safe and healthy space for students and parents to access customized sex education content recommendations, and peer story-sharing of lived experiences.

Website Figma

Hungry24

Hungry24 is a multi-puporse application hosted on Wechat mini-app platform that aims to encourage participants to accept a 24-hour hunger challenge. It's a collaborative fundraising effort with China Foundation for Poverty Alleviation to alleviate starvation for Ethiopian children.

Code PDF Article

FeedMe

Auto waste classification based on object detection.

FeedMe is a smart garbage classification and collection device developed at the X Academy 2018 hackathon. It was ranked the second place and won $10K funding.

Code PDF

Ph.D. Student @ University of Washington

Recent highlights

Selected publications

SugarCrepe: Fixing Hackable Benchmarks for Vision-Language Compositionality

CREPE: Can Vision-Language Foundation Models Reason Compositionally?

Model Sketching: Centering Concepts in Early-Stage Machine Learning Model Design

ELIGN: Expection Alignment as a Multi-agent Intrinsic Reward

Openattack: An Open-source Textual Adversarial Attack Toolkit

Awards