Siqi Sun
 |
Associate Professor from |
About Me
I am a tenure-track Associate Professor (Young PI) from Fudan University and Shanghai AI Lab. Previously, I was a Researcher from Knowledge and Language Team under Cognitive Services Research Group at Microsoft. I am mainly interested in deep learning and its applications in NLP and computational biology. More specifically, model compression/acceleration, knowledge graph, information retrieval, dialog response generation, protein/RNA structure prediction.
Before that, I was a PhD student at TTI-Chicago, a philanthropically endowed computer science research institute in University of Chicago . For more about TTIC, please refer to About TTIC. And I am fortunate to work with Prof. Jinbo Xu.
Prior to joining TTIC in 2011, I was a undergraduate student from School of Mathematics, Fudan University, China.
For more details, please refer to my LinkedIn.
Internships & Industrial Experiences
- Mar 2022 - present, Associate Professor, Fudan University
- Mar 2022 - present, - , Shanghai AI Lab
- Mar 2018 - Mar 2022, Researcher, Microsoft
- Jun 2015 - Sep 2015, Research Intern, Adobe Research , Mentor: Branislav Kveton, Hung Bui
- Mar 2011 - Jun 2011, Quant Research Intern, UBS Shanghai , Mentor: Wenfeng Cao
- Jul 2010 - Nov 2010, Quant Research Intern, Opera Solutions Shanghai , Mentor: Karl Xu
News
- Our HGN algorithm achieves NO.1 on HotpotQA for both distractor setting and fullwiki setting.
- Our DialoGPT project was featured by Adweek, The Register, Synced and Jiqizhixin(机器之心, in Chinese).
- My manager and I were interviewed by The New York Times for our NO.1 leaderboard results on both ARC and CommonsenseQA..
- Our FreeLB paper achieves NO.1 on GLUE benchmark and was reported by InfoQ.
- Our deep learning paper for contact prediction won the 2018 PLoS Computational Biology Research Prize in the category of Breakthrough/Innovation.
- I will join Microsoft(Redmond) in 2018!
- My work with Prof Jinbo Xu was reported by Science and The Economist .
- Our contact prediction algorithm won the 1st place in CASP12! Please see this paper for the official ranking table.
Publications
*-equal contributionJoint retrieval and generation training for grounded text generation
Yizhe Zhang, Siqi Sun, Xiang Gao, Yuwei Fang, Chris Brockett, Michel Galley, Jianfeng Gao, Bill Dolan
arxiv 2021
LightningDOT: Pre-training Visual-Semantic Embeddings for Real-Time Image-Text Retrieval
Siqi Sun*, Yen-Chun Chen*, Linjie Li, Shuohang Wang, Yuwei Fang, Jingjing Liu
NAACL 2021
Cluster-Former: Clustering-based Sparse Transformer for Long-Range Dependency Encoding
Shuohang Wang, Luowei Zhou, Zhe Gan, Yen-Chun Chen, Yuwei Fang, Siqi Sun, Yu Cheng, Jingjing Liu
Findings of ACL 2021
FILTER: An Enhanced Fusion Method for Cross-lingual Language Understanding
Yuwei Fang, Shuohang Wang, Zhe Gan, Siqi Sun, Jingjing Liu
AAAI 2021
Accelerating Real-Time Question Answering via Question Generation
Yuwei Fang, Shuohang Wang, Zhe Gan, Siqi Sun, Jingjing Liu
arxiv 2020
Cross-Thought for Sentence Encoder Pre-training
Shuohang Wang, Yuwei Fang, Siqi Sun, Zhe Gan, Yu Cheng, Jingjing Liu and Jing Jiang
EMNLP 2020
Contrastive Distillation on Intermediate Representations for Language Model Compression
Siqi Sun, Zhe Gan, Yu Cheng, Yuwei Fang, Shuohang Wang, Jingjing Liu
EMNLP 2020
Hierarchical Graph Network for Multi-hop Question Answering
Yuwei Fang, Siqi Sun, Zhe Gan, Rohit Pillai, Shuohang Wang, Jingjing Liu
EMNLP 2020
DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation
Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan
ACL 2020 Demo [Code]
FreeLB: Enhanced Adversarial Training for Language Understanding
Chen Zhu, Yu Cheng, Zhe Gan, Siqi Sun, Tom Goldstein, Jingjing Liu
ICLR 2020 (spotlight)
Patient Knowledge Distillation for BERT Model Compression
Siqi Sun, Yu Cheng, Zhe Gan, Jingjing Liu
EMNLP 2019 [Code]
Analysis of deep learning methods for blind protein contact prediction in CASP12
Sheng Wang, Siqi Sun and Jinbo Xu
Proteins: Structure, Function, and Bioinformatics, 2018
Accurate De Novo Prediction of Protein Contact Map by Ultra-Deep Learning Model
Sheng Wang*, Siqi Sun*, Zhen Li, Renyu Zhang, Jinbo Xu
PLOS Computational Biology, 2017 [Server]
Our server (RaptorX-Contact) was ranked 1st in contact prediction in the worldwide protein structure prediction (CASP) competition 12!
Learning Nonparametric Forest Graphical Models with Prior Information
Yuancheng Zhu, Zhe Liu and Siqi Sun
Artificial Intelligence and Statistics Conference (AISTATS' 2017)
Graphical Model Sketch
Branislav Kveton, Hung Bui, Mohammad Ghavamzadeh, Georgios Theocharous, S. Muthukrishnan and Siqi Sun
The European Conference on Machine Learning (ECML-PKDD' 2016)
AUC-maximized Deep Convolutional Neural Fields for Sequence Labeling
Sheng Wang, Siqi Sun and Jinbo Xu
The European Conference on Machine Learning (ECML-PKDD' 2016) [Code] [Server]
Learning structured densities via infinite dimensional exponential families
Siqi Sun, Mladen Kolar and Jinbo Xu
Neural Information Processing System (NIPS' 2015) [Supplementary Code]
Predicting diverse M-best protein contact maps
Siqi Sun, Jianzhu Ma, Sheng Wang and Jinbo Xu
IEEE International Conference on Bioinformatics and Biomedicine (BIBM' 2015)
Inferring Block Structure of Graphical Models in Exponential Families
Siqi Sun*, Hai Wang* and Jinbo Xu
Artificial Intelligence and Statistics Conference (AISTATS' 2015)
Learning Scale-Free Networks by Dynamic Node Specific Degree Prior
Qingming Tang*, Siqi Sun* and Jinbo Xu
International Conference on Machine Learning (ICML' 2015) [Code]
Adaptive Variable Clustering in Gaussian Graphical Models
Siqi Sun*, Yuancheng Zhu* and Jinbo Xu
Artificial Intelligence and Statistics Conference (AISTATS' 2014) [Code]
An iterative network partition algorithm for accurate identification of dense network modules
Siqi Sun*, Xinran Dong*, Yao Fu and Weidong Tian
Nucleic Acids Research 2011
Professional Activities
Reviewer of AISTATS 2016 - 2018, ECML 2016 - 2018, ICML 2015, NIPS 2015-2019, IEEE/ACM TCBB
Program Committee of ECML-PKDD 2017 and 2018 for the Applied Data Science Track