Jing Nathan Yan
I am a Research Scientist at Google DeepMind, where I explore model- and data-efficient approaches to Large Language Models.
I received my Ph.D. from Cornell CIS and Cornell Tech, where I was extremely fortunate to be advised by Prof. Sasha Rush and Prof. Jeff Rzeszotarski. Before coming to Cornell, I was affiliated with HKU (courses), working with Prof. Reynold Cheng, and also spent some time at SFU working with Prof. Jiannan Wang.
During my doctoral studies I had the great fortunate to spend summers working with fantastics researchers at Google Research, Microsoft Research and Facebook AI.
My research interest has always been drawn to the question, “How far can we stretch a bit of computation?”.
Here is my Email: jy858 AT cornell.edu
Selected Recent Publications[Full List]
-
Jing Nathan Yan*, Jiatao Gu*, Alexander Rush
Diffusion Models Without Attention.
In CVPR 2024, Jun 17 - 21, Seattle, United States. (*equal contribution)
A shorter version at NeurIPS 2023 Workshop on Diffusion Models -
Jing Nathan Yan, Tianqi Liu, Justin T. Chiu, Jiaming Shen, Zhen Qin, Yue Yu, Yao Zhao, Charu Lakshmanan, Yair Kurzion, Alexander Rush, Jialu Liu, Michael Bendersky
Predicting Text Preference Via Structured Self-Comparative Reasoning.
In ACL 2024, August 11 – 16, Bangkok, Thailand. [arXiv link] -
Yi Lu*, Jing Nathan Yan*, Songlin Yang, Justin Chiu, Siyu Ren, Fei Yuan, Wenting Zhao, Zhiyong Wu, Alexander Rush
A Controlled Study on Long-Context Extension and Generalization in LLMs.
COLM 2025, October 7-10, Montreal, Canada. (*equal contribution) -
Woojeong Kim, Junxiong Wang, Jing Nathan Yan, Mohamed S. Abdelfattah, Alexander Rush
Overfill: Two-Stage Models for Efficient Language Model Decoding.
COLM 2025, October 7-10, Montreal, Canada. -
Jing Nathan Yan, Junxiong Wang, Jeff Rzeszotarski, Allison Koenecke
Fairness Practices in Industry: A Case Study in Machine Learning Teams Building Recommender Systems.
Preprint 2025. -
Qi Sun, Zhiyang Guo, Ziyu Wan, Jing Nathan Yan, Shengming Yin, Wengang Zhou, Jing Liao, Houqiang Li
EG4D: Explicit Generation of 4D Object without Score Distillation.
ICLR 2025, Apr 24 – 28, Singapore. -
Yue Yu, Jiaming Shen, Tianqi Liu, Zhen Qin, Jing Nathan Yan, Jialu Liu, Chao Zhang, Michael Bendersky
Explanation-aware Soft Ensemble Empowers Large Language Model In-context Learning.
In ACL 2024, August 11 – 16, Bangkok, Thailand. [arXiv link] -
Junxiong Wang, Tushaar Gangavarapu, Jing Nathan Yan, Alexander Rush
MambaByte: Token-free Selective State Space Model.
Preprint 2024. [arXiv link] -
Junxiong Wang, Jing Nathan Yan, Albert Gu, Alexander Rush
Pretraining Without Attention.
In EMNLP 2023 Findings, Dec 5 - 10, Singapore. -
Jing Nathan Yan, Ziwei Gu, Jeff Rzeszotarski
Tessera: Discretizing Data Analysis Workflows on a Task Level.
In CHI 2021, May 8 - 13, Yokohama, Japan. -
Ziwei Gu*, Jing Nathan Yan*, Jeff Rzeszotarski
Understanding User Sensemaking in Machine Learning Fairness Assessment Systems.
In TheWeb 2021(WWW 2021), April 19 - 23, Ljubljana, Slovenia(*equal contribution). [Datasets] -
Jing Nathan Yan, Oliver Schulte, Mohan Zhang, Jiannan Wang, Reynold Cheng
SCODED: Statistical Constraint Oriented Data Error Detection.
In SIGMOD 2020, June 14 - 19, Portland, United States. -
Jing Nathan Yan, Ziwei Gu, Hubert Lin, Jeff Rzeszotarski
Silva: Interactively Assessing Machine Learning Fairness Using Causality.
In CHI 2020, April 25 - 30, Hawai'i, United States.
Professional Activities
Program Committee
- FaccT 2023, 2024; AIES 2025
Reviewers
- ICLR, NeurIPS, ARR, CHI, CSCW, UIST, SIGMOD, VLDB, KDD, ICDE, TKDE, VIS, SoCC, CIKM
Personal
Reading books(The GAY GENIUS SU SHI), Doing random Weight lifting, and Trying to practice golfing more. Also a big fan of many Japanese Cartoons 名探偵コナン and 夏目友人帳.
Was an active member in Foreign Affairs Association. Co-founder of PowerFair, a NGO platform based in Hong Kong which aims at promoting the equal access to public resources (i.e., data).
Previous band member, good at making noises.