Yufan Zhuang
Anywhere on campus :)
9500 Gilman Drive
La Jolla, California
Hey, I’m Yufan Zhuang (庄宇凡)!
I am a PhD student in UCSD’s Computer Science & Engineering department advised by Jingbo Shang. My research is centered on natural language processing and large language models.
I have proposed ways to in-context learn from any modality (Vector-ICL), conducted studies on cross-lingual data contamination for contemporary LLMs, made transformers effective meta algorithm learners for decision trees (MetaTree).
I’ve also trained first series of high quality Mamba-based vision language models (ViperVLMs), extending LLMs context length with adaptive wavelet transform (WavSpa), and a series of work on computational sociology research.
Prior to my PhD study, I worked at IBM T. J. Watson Research Center as a research engineer helping to enhance software engineering with the power of AI and vice versa. I received my MS in Data Science from Columbia, my BSc in Applied Math Minor in CS (with First Class Honors) from Hong Kong Polytechnic University.
news
Sep 01, 2024 | Our paper Data Contamination Can Cross Language Barriers has been published at EMNLP 2024! |
---|---|
Aug 30, 2024 | Our paper Learning a Decision Tree Algorithm with Transformers has been published at Transactions on Machine Learning Research! |
Jun 19, 2024 | I’ve started my summer internship at the Meta Reality Lab (Menlo Park, California), where we developed first series of high-quality mamba-based vision LLMs (Viper-Mamba-7B and Viper-Jamba-52B) |
Jun 19, 2023 | I’ve started my summer internship at the Deep Learning group of Microsoft Research (Redmond, Washington), mentored by Chandan Singh and Liyuan Liu! |
May 12, 2023 | Our paper “Incorporating Signal Awareness in Source Code Modeling: an Application to Vulnerability Detection” has been published at ACM Transactions on Software Engineering and Methodology! |
selected publications
- arXivVector-ICL: In-context Learning with Continuous Vector RepresentationsarXiv preprint arXiv:2410.05629, 2024
- EMNLPData Contamination Can Cross Language BarriersEmpirical Methods in Natural Language Processing, 2024
- TMLRLearning a Decision Tree Algorithm with TransformersTransactions on Machine Learning Research, 2024
- NeurIPSWavSpA: Wavelet Space Attention for Boosting Transformers’ Long Sequence Learning AbilityNeurIPS 1st UniReps Workshop, 2023
- TSEIncorporating Signal Awareness in Source Code Modeling: an Application to Vulnerability DetectionACM Transactions on Software Engineering and Methodology, 2023
- AAGSleeping Lion or Sick Man? Machine Learning Approaches to Deciphering Heterogeneous Images of Chinese in North AmericaAnnals of the American Association of Geographers, 2022
- FSEProbing model signal-awareness via prediction-preserving input minimizationACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering, 2021
- BDR