Xingchen Wan

Research Scientist

Google

About me

I am a Research Scientist at Google based in the San Francisco Bay Area.

I did my DPhil (the Oxford way of saying PhD) in the Machine Learning Research Group, Department of Engineering Science, University of Oxford, where I was supervised by Professor Michael A. Osborne. I was also a Clarendon Scholar and a member of St John’s College, both at the University of Oxford. I previously interned at Google and Meta.

Recent News

Feb 2024: I started my position as a Research Scientist at Google!
Jan 2024: 3 paper acceptances: Gadam accepted at JMLR [preprint]; AutoPEFT accepted at TACL [preprint]; AdaBatAL accepted at AISTATS 2024 [preprint].
Jan 2024: Our work, Batch Calibration, was accepted at ICLR 2024. The work was featured in the Google Research blogs, and an earlier version appeared as a Spotlight presentation at the Workshop on Robustness of Few-shot and Zero-shot Learning in Foundation Models @ NeurIPS 2023.
Dec 2023: Our work analyzing the working memory of LLMs like ChatGPT was accepted at AAAI 2024.
Nov 2023: I was recognized as a top reviewer at NeurIPS 2023.
Nov 2023: I passed my PhD viva voce a.k.a. oral defence!
Nov 2023: COSP and USP are featured in a recent Google Research blog!
Oct 2023: Two papers were accepted at EMNLP: Universal Self-Adaptive Prompting; CLaPS.
Sep 2023: Our work on Bayesian optimization on graphs was accepted at NeurIPS.

Academic Services

Reviewer/program committee member at ACL (2023-24), ICML (2023-24), AutoML-Conf (2023-24), CVPR (2024), ECCV (2024), ICLR (2024), WACV (2022-24), NeurIPS (2022-23), EMNLP (2023), JMLR, Machine Learning, etc.

Area chair at NeurIPS (2024).

Interests

Large language models
Bayesian optimization & AutoML
Machine learning on graphs and networks

Education

DPhil (PhD) in Machine Learning, 2019 - 2023

University of Oxford
MEng (integrated bachelor's and master's degree) in Engineering Science. 1st-class Honours (top graduate of the class), 2015 - 2019

University of Oxford

Featured

Batch Calibration: Rethinking Calibration for In-Context Learning and Prompt Engineering

Mitigating prompt biases and unifying existing calibration approaches without labeled data (ICLR 2024)

Han Zhou, Xingchen Wan, Lev Proleev, Diana Mincu, Jilin Chen, Katherine Heller, Subhrajit Roy

Batch Calibration: Rethinking Calibration for In-Context Learning and Prompt Engineering

AutoPEFT: Automatic Configuration Search for Parameter-Efficient Fine-Tuning

Automatic discovery of families of high-performing PEFT configurations (TACL 2024)

Han Zhou, Xingchen Wan, Ivan Vulić, Anna Korhonen

AutoPEFT: Automatic Configuration Search for Parameter-Efficient Fine-Tuning

Survival of the Most Influential Prompts: Efficient Black-Box Prompt Search via Clustering and Pruning

Orders-of-magnitude faster hard prompt search with SoTA performance (EMNLP Findings 2023)

Han Zhou, Xingchen Wan, Ivan Vulić, Anna Korhonen

Introducing Self-Adaptive Prompting for Large Language Models

Adaptive, black-box and automatic prompting methods to improve LLMs in zero-shot settings (ACL & EMNLP 2023)

Xingchen Wan

Last updated on 24 Oct 2023 8 min read

Introducing Self-Adaptive Prompting for Large Language Models

Interpretable Neural Architecture Search via Bayesian Optimisation with Weisfeiler-Lehman Kernels

Performant and efficient NAS method combining BO and graph kernels (ICLR 2021)

Binxin Ru, Xingchen Wan, Xiaowen Dong, Michael Osborne

Interpretable Neural Architecture Search via Bayesian Optimisation with Weisfeiler-Lehman Kernels

Publications

View all listed publications or view by tags. For a complete list including preprints & working papers, refer to my Google Scholar.

Han Zhou, Xingchen Wan, Lev Proleev, Diana Mincu, Jilin Chen, Katherine Heller, Subhrajit Roy (2024). Batch Calibration: Rethinking Calibration for In-Context Learning and Prompt Engineering. International Conference on Learning Representations (ICLR).

PDF Cite Google Research Blog Abstract OpenReview

Masaki Adachi, Satoshi Hayakawa, Martin Jørgensen, Xingchen Wan, Vu Nguyen, Harald Oberhauser, Michael Osborne (2024). Adaptive Batch Sizes for Active Learning: A Probabilistic Numerics Approach. International Conference on Artificial Intelligence and Statistics (AISTATS).

PDF Cite Abstract

Dongyu Gong, Xingchen Wan, Dingmin Wang (2024). Working Memory Capacity of ChatGPT: An Empirical Study. AAAI Conference on Artificial Intelligence (AAAI).

PDF Cite Code Abstract

Han Zhou, Xingchen Wan, Ivan Vulić, Anna Korhonen (2024). AutoPEFT: Automatic Configuration Search for Parameter-Efficient Fine-Tuning. Transactions of the Association for Computational Linguistics (TACL).

PDF Cite Code Abstract

Diego Granziol, Nick P. Baskerville, Xingchen Wan, Samuel Albanie, Stephen Roberts (2024). Iterate Averaging in the Quest for Best Test Error. Journal of Machine Learning Research (JMLR).

PDF Cite Code Abstract

Xingchen Wan, Ruoxi Sun, Hootan Nakhost, Hanjun Dai, Julian Martin Eisenschlos, Sercan Ö. Arık, Tomas Pfister (2023). Universal Self-Adaptive Prompting. Empirical Methods in Natural Language Processing (EMNLP).

PDF Cite Google Research Blog Abstract (Google Research) OpenReview ACL Anthology

Han Zhou, Xingchen Wan, Ivan Vulić, Anna Korhonen (2023). Survival of the Most Influential Prompts: Efficient Black-Box Prompt Search via Clustering and Pruning. Findings of the Association for Computational Linguistics: EMNLP 2023.

PDF Cite Code Abstract OpenReview ACL Anthology

Xingchen Wan, Pierre Osselin, Henry Kenlay, Binxin Ru, Michael Osborne, Xiaowen Dong (2023). Bayesian Optimisation of Functions on Graphs. Advances in Neural Information Processing Systems (NeurIPS).

PDF Cite Abstract OpenReview

See all publications