Vision and Language

• Pengchuan Zhang*, Xiujun Li*, Xiaowei Hu, Jianwei Yang, Lei Zhang, Lijuan Wang, Yejin Choi, Jianfeng Gao
  VinVL: Making Visual Representations Matter in Vision-Language Models, CVPR 2021
  Impact: OSCAR+ is the current SoTA on seven downstream tasks, and also No.1 on three public leaderboards (VQA, Image Captioning, and NoCaps).
• Xiujun Li, Xi Yin, Chunyuan Li, Xiaowei Hu, Pengchuan Zhang, Lei Zhang, Lijuan Wang, Houdong Hu, Li Dong, Furu Wei, Yejin Choi, Jianfeng Gao
  Oscar: Object-Semantics Aligned Pre-training for Vision-Language Tasks, ECCV 2020
  Impact: Oscar is widely used in several Microsoft products.
• Chunyuan Li, Xiujun Li, Lei Zhang, Baolin Peng, Jianfeng Gao
  Self-supervised Pre-training with Hard Examples Improves Visual Representations, 2020
  Impact: SoTA (75.5% Top1 accuracy) on linear classification on ImageNet.

Dialog, RL

• Xiujun Li, Yun-Nung Chen, Lihong Li, Jianfeng Gao, Asli Celikyilmaz
  End-to-End Task-Completion Neural Dialogue Systems, IJCNLP 2017
• Baolin Peng, Xiujun Li, Lihong Li, Jianfeng Gao, Asli Celikyilmaz, Sungjin Lee, Kam-Fai Wong
  Composite Task-Completion Dialogue System via Hierarchical Deep Reinforcement Learning, EMNLP 2017
• Baolin Peng, Xiujun Li, Jianfeng Gao, Jingjing Liu, Kam-Fai Wong and Shang-Yu Su
  Deep Dyna-Q: Integrating Planning for Task-Completion Dialogue Policy Learning, ACL 2018