I am a research scientist at Google Brain on Dale Schuurmans’s team and a PhD student at UC Berkeley advised by Pieter Abbeel. I am broadly interested in reinforcement learning and representation learning for sequential decision making.

Prior to Brain, I received my M.Eng. and B.S. in computer science from MIT working with Julian Shun on high-performance computing for large-scale graphs. I have also worked on interesting computer systems problems such as multi-core inference in TensorFlow TPU and lazy memory allocation in the Android operating system.

 

Preprints

  • Mengjiao Yang, Dale Schuurmans, Pieter Abbeel, Ofir Nachum. Chain of Thought Imitation Learning with Procedure Cloning. [paper][code][slides][website]

Selected Publications

  • Mengjiao Yang, Sergey Levine, Ofir Nachum. TRAIL: Near-Optimal Imitation Learning with Suboptimal Data. ICLR, 2022. [paper][code][slides][video][website]
  • Mengjiao Yang*, Bo Dai*, Ofir Nachum*, George Tucker, Dale Schuurmans. Offline Policy Selection under Uncertainty. AISTATS, 2022. [paper][code][slides][video]
  • Ofir Nachum, Mengjiao Yang. Provable Representation Learning for Imitation with Contrastive Fourier Features. NeurIPS, 2021. [paper][code]
  • Mengjiao Yang, Ofir Nachum. Representation Matters: Offline Pretraining for Sequential Decision Making. ICML, 2021. [paper][code][slides]
  • Mengjiao Yang*, Ofir Nachum*, Bo Dai*, Lihong Li, Dale Schuurmans. Off-Policy Evaluation via the Regularized Lagrangian. NeurIPS, 2020. [paper][code][slides]
  • Mengjiao Yang*, Bo Dai*, Hanjun Dai, Dale Schuurmans. Energy-Based Processes for Exchangeable Data. ICML, 2020. [paper][code][slides][video]
  • Mengjiao Yang, Been Kim. Benchmarking Attribution Methods with Relative Feature Importance. NeurIPS workshop (oral :sparkles:) on Human-Centric Machine Learning, 2019. [paper][code][poster]

Other Publications

  • Siddharth Verma, Justin Fu, Mengjiao Yang, Sergey Levine. CHAI: A CHatbot AI for Task-Oriented Dialogue with Offline Reinforcement Learning. NAACL, 2022 (oral :sparkles:). [paper][website]
  • Charlie Snell, Mengjiao Yang, Justin Fu, Yi Su, Sergey Levine. Context-Aware Language Modeling for Goal-Oriented Dialogue Systems. NAACL, 2022. [paper][code][website]
  • Hongyu Ren, Hanjun Dai, Zihang Dai, Mengjiao Yang, Jure Leskovec, Dale Schuurmans, Bo Dai. Combiner: Full Attention Transformer with Sparse Computation Cost. NeurIPS, 2021 (spotlight :sparkles:). [paper][code]
  • Haoming Jiang, Bo Dai, Mengjiao Yang, Tuo Zhao, Wei Wei. Towards Automatic Evaluation of Dialog Systems: A Model-Free Off-Policy Evaluation Approach. EMNLP, 2021. [paper][code]
  • Justin Fu, Mohammad Norouzi, Ofir Nachum, George Tucker, Ziyu Wang, Alexander Novikov, Mengjiao Yang, R. Michael Zhang, Yutian Chen, Aviral Kumar, Cosmin Paduraru, Sergey Levine, Thomas Paine. Benchmarks for Deep Off-Policy Evaluation. ICLR, 2021. [paper][code]
  • Mengjiao Yang. Cache and NUMA Optimizations in A Domain-Specific Language for Graph Processing. MIT M.Eng. Thesis, 2018. [paper][slides]
  • Yunming Zhang, Mengjiao Yang, Riyadh Baghdadi, Shoaib Kamil, Julian Shun, Saman Amarasinghe. GraphIt: A High-Performance Graph DSL. OOPSLA, 2018. [paper][code][slides][website]