Selected Works

• DiJia Andy Su, Hanlin Zhu, Yingchen Xu *, Jiantao Jiao, Yuandong Tian^, Qinqing Zheng^.

  Token Assorted: Mixing Latent and Text Tokens for Improved Language Model Reasoning

  Preprint. Arxiv.

• DiJia Andy Su, Andrew Gu, Jane Xu, Yuandong Tian, Jiawei Zhao

  Galore 2: Large Scale Low Rank Memory Efficient LLM Pretraining on 500B tokens

  Preprint. Arxiv. (Large scale LLM pretraining)

• Shibo Hao, Sainbayar Sukhbaatar, DiJia Andy Su, Xian Li, Zhiting Hu, Jason Weston, Yuandong Tian

  Training Large Language Models to Reason in a Continuous Latent Space

  Preprint. Arxiv.
  
  

• [1] DiJia Andy Su, Sainbayar Sukhbaatar, Michael Rabbat, Yuandong Tian, Qinqing Zheng

  Dualformer: Controllable Fast and Slow Thinking by Learning with Randomized Reasoning Traces

  International Conference on Learning Representations (ICLR) 2025.
  
  

• [2] Lucas Lehnert, Sainbayar Sukhbaatar, DiJia Andy Su, Qinqing Zheng, Paul Mcvay, Michael Rabbat, Yuandong Tian

  Beyond A*: Better Planning with Transformers via Search Dynamics Bootstrapping

  Conference of Language Modeling (COLM) 2024

• [3] DiJia Andy Su,  Jayden Ooi, Tyler Lu, Dale Schuurmans, Craig Boutilier 

        Conqur: Mitigating Delusional Bias in Deep Q-Learning

       International Conference on Machine Learning (ICML) 2021

- Note: Followup work on Non-delusional Q Learning (NeurIPS 18 Best Paper)

• [4] DiJia Andy Su, Bertrand Douillard, Rami Al-Rfou, Cheol Park, Benjamin Sapp

        Narrowing the Coordinate-Frame Gap in Behavior Prediction Models: Distillation for efficient and accurate scene-centric motion forecasting (training 15B model) 

         International Conference on Robotics and Automation (ICRA) 2022

         - Note: Won 1st Place at Open Motion Prediction Challenge, as of Sep 1

  
  

• [5] DiJia Andy Su, Jason D Lee, John M Mulvey, H Vincent Poor

  Competitive Multi-Agent Reinforcement Learning with Self-Supervised Representation

        IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2022

• [6] DiJia Andy Su, Jason D Lee, John M Mulvey, H Vincent Poor

  MUSBO: Model-based uncertainty regularized and sample efficient batch optimization for deployment constrained reinforcement learning

         Under minor revision at Journal of Machine Learning Research (JMLR).  

• [7] DiJia Andy Su, DF Su, Jason D Lee, John M Mulvey, H Vincent Poor

  Optimizing Multi-Document Summarization by Blending Reinforcement Learning Policies

        IEEE Transactions on Artificial Intelligence

• [8] DiJia Andy Su, John M Mulvey, H Vincent Poor

  Improving Portfolio Performance via Natural Language Processing Methods

        Journal of Financial Data Science

• [9] DiJia Andy Su *, Zihan Ding*, Qinghua Liu, Chi Jin 

       A Deep Reinforcement Learning Approach for Finding Non-Exploitable Strategies in Two-Player Atari Games

       Arxiv 2022.

• [10] DiJia Andy Su, et al.

  Machine Learning Optimization with Crypto Currency

        Under Revision at Operation Research

• [11] DiJia Andy Su, SA Mousavifar, Cyril Leung

  Secrecy capacity and wireless energy harvesting in amplify-and-forward relay networks

        IEEE Pacific Rim Conference on Communications, Computers and Signal Processing

        Note: Won best paper award. 

• [12] Yunfei Zhang, Clarence W de Silva, DiJia Andy Su, Youtai Xue

  Autonomous robot navigation with self-learning for collision avoidance with randomly moving obstacles

        International Conference on Computer Science (ICCSE)

        Note: Won best paper award. 

Patents:

• Tyler Lu, Jayden Ooi, Craig Boutilier, Dale Schuurmans, DiJia Andy Su

  Mitigating delusional bias in deep q-learning for robotic and/or other agents

   Patent Number: US20220101111A1

• Bertrand Douillard, DiJia Andy Su

  Training agent trajectory prediction neural networks using distillation

   Patent Number: US20230082079A1