Pengfei Yu

Pengfei Yu

PhD Student in Computer Science

University of Illinois, Urbana-Champaign

Biography

My Chinese name is 于(Yu)鹏(Peng)飞(fei). I am a PhD student in computer science at UIUC blender lab advised by Heng Ji. My research interests include information extraction (IE) and its applications in NLP. I also have some experience on a variety of topics, such as meta learning, transfer learning, continual learning, weakly-supervised learning and multimodal learning.

Interests
  • Natural Langauge Processing
  • Machine Learning
  • Information Extraction
Education

  • PhD in Computer Science, 2019-Now

    University of Illinois, Urbana-Champaign

  • BSc in Mathematics and Applid Mathematics, 2016-2019

    Tsinghua University

  • BEng in Electronic Engineering, 2015-2019

    Tsinghua University

Research Experience

 
 
 
 
 
UIUC Blender Lab, with Prof. Heng Ji
Research Assistant
UIUC Blender Lab, with Prof. Heng Ji
Aug 2019 – Present Illinois, US
Researched in information extraction and NLP
 
 
 
 
 
Amazon Alexa AI
Applied Scientist Intern
Amazon Alexa AI
May 2022 – Aug 2022 California, US
Research in knowledge graph and information extraction
 
 
 
 
 
UIUC
Teaching Assistant
UIUC
Aug 2021 – Dec 2021 Illinois, US
TA for Advanced Topics in NLP
 
 
 
 
 
UCSB NLP Group, with Prof. William Yang Wang
Research Assistant
UCSB NLP Group, with Prof. William Yang Wang
Jun 2018 – Aug 2018 California, US
Researched in knowledge representation
 
 
 
 
 
Tsinghua THUNLP, with Prof. Zhiyuan Liu
Research Assistant
Tsinghua THUNLP, with Prof. Zhiyuan Liu
Sep 2017 – Jul 2019 Beijing, China
Researched in relation extraction and knowledge representation

Recent Posts

Updated Thoughts on Randomized Language Models
This is a updated version of my previous post on stochastic language models. This post can be read alone, since I have modified the previous idea significantly. Simply put, I still believe a stochastic language model is more desirable than a constant language model, but my previous immature post is subject to a few fallacies.
Pengfei Yu
Apr 12, 2024 9 min read Ideas
Thoughts on Randomizing Lanugage Modeling for Better Language Learning
Considering language models as random variables brings insights into the challenges in LLMs and the mechanisms of various augmentation methods to improve LLMs on hallucination, biases, continual fine-tuning/knowledge editing/instruction fine-tuning/RLHF, tool learning, role playing, retrieval augmentation…
Pengfei Yu
Oct 25, 2023 17 min read Ideas

Featured Publications

Pengfei Yu, Jiateng Liu, Pengfei Yu,, Yuji Zhang, Sha Li, Zixuan Zhang, Heng Ji
February, 2024
EVEDIT: Event-based Knowledge Editing with Deductive Editing Boundaries

Counterfactual triple-based knowledge editing is subject to logical fallacies due to a lack of deduction anchors. We fix it by introducing EVent-based Editing.
PDF Code Dataset Poster Video Source Document arxiv
Pengfei Yu, Heng Ji
May, 2023
Information Association for Language Model Updating by Mitigating LM-Logical Discrepancy

Language modeling doesn’t imply information modeling. Optimzing LM probabilities is not enough to keep LLMs updated.
PDF Code Dataset Poster Video Source Document arxiv
Pengfei Yu, Heng Ji, Prem Natarajan
November, 2021 Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing
Lifelong Event Detection with Knowledge Transfer

Build an event detection model that expands its ontology dynamically forever.
PDF Cite DOI

Publications

Pengfei Yu, Jiateng Liu, Pengfei Yu,, Yuji Zhang, Sha Li, Zixuan Zhang, Heng Ji (2024). EVEDIT: Event-based Knowledge Editing with Deductive Editing Boundaries.

PDF Code Dataset Poster Video Source Document arxiv

Pengfei Yu, Heng Ji (2023). Information Association for Language Model Updating by Mitigating LM-Logical Discrepancy.

PDF Code Dataset Poster Video Source Document arxiv

Pengfei Yu, Zixuan Zhang, Clare Voss, Jonathan May, Heng Ji (2022). Event Extractor with Only a Few Examples. Proceedings of the 3rd Workshop on Deep Learning Approaches for Low-Resource NLP (DeepLo 2022).

Cite

Xinya Du, Zixuan Zhang, Sha Li, Pengfei Yu, Hongwei Wang, Tuan Manh, Xudong Lin, Ziqi Wang, Iris Liu, Ben Zhou, Haoyang Wen, Manling Li, Darryl Hannan, Qi Zeng, Qing Lyu, Charles Yu, Carl Edwards, Xiaomeng Jin, Yizhu Jiao, Ghazaleh Kazeminejad, Rotem Dror, Zhenhailong Wang, Chris Callison-Burch, Mohit Bansal, Carl Vondrick, Jiawei Han, Dan Roth, Shih-Fu Chang, Martha Palmer, Heng Ji (2022). RESIN-11: Schema-guided Event Prediction for 11 Newsworthy Scenarios. Proceedings of 2022 Annual Conference of the North American Chapter of the Association for Computational Linguistics: System Demonstrations.

Cite

Manling Li, Revanth Gangi Reddy, Ziqi Wang, Yi-shyuan Chiang, Tuan Lai, Pengfei Yu, Zixuan Zhang, Heng Ji (2022). COVID-19 Claim Radar: A Structured Claim Extraction and Tracking System. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics: System Demonstrations.

PDF Cite DOI

Pengfei Yu, Heng Ji, Prem Natarajan (2021). Lifelong Event Detection with Knowledge Transfer. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing.

PDF Cite DOI

Xiaodan Hu, Pengfei Yu, Kevin Knight, Heng Ji, Bo Li, Honghui Shi (2021). MUSE: Textual Attributes Guided Portrait Painting Generation. 2021 IEEE 4th International Conference on Multimedia Information Processing and Retrieval (MIPR).

PDF Cite DOI

Xu Han, Hao Zhu, Pengfei Yu, Ziyun Wang, Yuan Yao, Zhiyuan Liu, Maosong Sun (2018). FewRel: A Large-Scale Supervised Few-Shot Relation Classification Dataset with State-of-the-Art Evaluation. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing.

PDF Cite DOI

Xu Han, Pengfei Yu, Zhiyuan Liu, Maosong Sun, Peng Li (2018). Hierarchical Relation Extraction with Coarse-to-Fine Grained Attention. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing.

PDF Cite DOI

Yinpei Dai, Zhijian Ou, Dawei Ren, Pengfei Yu (2018). Tracking of Enriched Dialog States for Flexible Conversational Information Access. 2018 IEEE International Conference on Acoustics, Speech and Signal Processing.

PDF Cite DOI

Hobbies

Sports

I am a big soccer fan and a supporter of Real Madrid since 11 years old. I was a fan of Kaká before he retired. I like his playing style, but I am not tall enough to run like him 🙃. In addition to soccer, I am also interested in most physical activities.

Movies and Video Games

I enjoy fantasy movies, a fan of Marvel, DC, Harry Potter, Star Wars, The Lord of the Rings… I also watch other categories.

I play video games, the Witcher III for 500 hours and ELden Ring for 250 hours 😂. I also play sport games such as FIFA and NBA 2K.

Funny Things

I am a playing cards collector and a very-beginner in card magics and cardistry (what’s cardistry?) too. I also like puzzles and Rubik’s cube, but not good at all.

Contact