Short Bio

I am an applied scientist at Amazon, working on the post-training stage of Rufus, the shopping LLM for Amazon.com. I got my Ph.D. at School of Computer Science, University of Illinois at Urbana-Champaign, advised by Prof. Heng Ji. Before coming to UIUC, I got my bachelor degree from Shanghai Jiao Tong University, where I was a member of Zhiyuan's Honors Program of Engineering.

My research mainly lies in building Knowledge-enhanced Language Models. Recently, I am particularly interested in making the knowledge enhancements to LLMs more generalizable, by enabling automatic knowledge extraction, disentangled knowledge storage, and consistent knowledge updates. I have also worked on other fundamental NLP tasks like information extraction, knowledge graph completion, question answering, and text summarization.

Here are my Google Scholar Profile and CV.

What's new:

2024.03 - Two papers accepted by EMNLP 2024.

2024.03 - Our paper on improving the generalizability of retrieval-augmented language models is accepted to NAACL 2024 Findings.

2023.08 - Finished my internship at Google Research.

Area of Interests

Currently, my research interests include:

  • Knowledge-enhanced Large Language Models
  • Language Models as Proactive Agents
  • Cross-document Information Extraction, Representation, and Understanding

Contact

Address Room 1111, Thomas M. Siebel Center,
201 N. Goodwin Avenue,
Urbana, IL 61801 USA.
Email zixuan11@illinois.edu
zhangzx.uiuc@gmail.com
​ ​