Scholar
Zhengyan Shi
Google Scholar ID: TF8l2ZEAAAAJ
Microsoft Research
Citations & Impact
Citations
430
H-index
10
i10-index
11
Publications
20
Co-authors
18
list available
Publications
Resume (English only)
Academic Achievements
- - Publications:
- - Understanding Likelihood Over-optimisation in Direct Alignment Algorithms, Preprint, 2024
- - Instruction Tuning With Loss Over Instructions, NeurIPS, 2024
- - DePT: Decomposed Prompt Tuning for Parameter-Efficient Fine-tuning, ICLR, 2024
- - Don't Stop Pretraining? Make Prompt-based Fine-tuning Powerful Learner, NeurIPS, 2023
- - Rethinking Semi-supervised Learning with Language Models, Findings of ACL, 2023
Research Experience
- - Microsoft Research: Current research focus is on teaching language models to code
- - Cohere (London): Research Intern
- - Amazon (London & Seattle): Research Intern
Education
- - PhD in Computer Science from University College London (UCL)
- - MSc in Data Science (Statistics) with Distinction from UCL
- - BSc in Mathematics with First Class Honours from the University of Liverpool and Xi'an Jiaotong-Liverpool University
Background
- I am a Senior Researcher at Microsoft Research (MSR). My current research focuses on teaching language models (LMs) to code. I build learning loops in which LMs not only act but also reason within scalable, self-evolving environments.
Miscellany
- Links: Google Scholar / Twitter / Github / LinkedIn / Email
