Kaixiang Lin

Senior applied scientist at Amazon AGI.

I am currently working on MultiModal RLHF for Amazon Nova Models. From Oct 2021 to Feb 2024, I worked on pre-training and contrastive learning of text embedding model, initiated and led RLHF efforts for Amazon Tian models at AWS. Before joining AWS, I spent one year at Amazon Lab126, working on embodied AI. Previously, I completed my Ph.D. in computer science at MSU and my B.S. in electronic engineering and information science at USTC.

Email  /  CV  /  Google Scholar  /  Twitter  /  Github

profile photo
News
Publications


2025

Do LLMs Recognize Your Preferences? Evaluating Personalized Preference Following in LLMs
Siyan Zhao, Mingyi Hong, Yang Liu, Devamanyu Hazarika, Kaixiang Lin
PDF / Code / Data / Website / ICLR 2025 (Oral)

2024

Proposer-Agent-Evaluator (PAE): Autonomous Skill Discovery For Foundation Model Internet Agents
Yifei Zhou*, Qianlan Yang*, Kaixiang Lin, Min Bai, Xiong Zhou, Yu-Xiong Wang, Sergey Levine, Erran Li
PDF / Code / Website/ Model checkpoints/* denotes equal contribution.

LLM Alignment Through Successive Policy Re-weighting (SPR)
Xinnan Zhang, Siliang Zeng, Jiaxiang Li, Kaixiang Lin, Mingyi Hong.
PDF / NeurIPS 2024 Workshop on Fine-Tuning in Modern Machine Learning: Principles and Scalability.

Socratic human feedback (SoHF): Expert steering strategies for LLM code generation
Subramanian Chidambaram*, Erran Li*, Min Bai, Xiaopeng LI, Kaixiang Lin, Xiong Zhou, Alex C. Williams.
PDF / EMNLP findings. /∗ denotes equal contribution

Bootstrapping LLM-based Task-Oriented Dialogue Agents via Self-Talk
Dennis Ulmer, Elman Mansimov, Kaixiang Lin, Justin Sun, Xibin Gao, Yi Zhang .
PDF

2023

Automated Few-Shot Classification with Instruction-Finetuned Language Models
Rami Aly, Xingjian Shi, Kaixiang Lin, Aston Zhang, Andrew Gordon Wilson.
PDF / EMNLP Findings .

A Unified Linear Speedup Analysis of Stochastic FedAvg and Nesterov Accelerated FedAvg
Zhaonan Qu*, Kaixiang Lin*, Zhaojian Li, Jiayu Zhou, Zhengyuan Zhou.
PDF / Journal of Artificial Intelligence Research (JAIR) / * denotes equal contribution.

Transfer Learning in Deep Reinforcement Learning: A Survey
Zhuangdi Zhu, Kaixiang Lin, Jiayu Zhou.
PDF / Transactions on Pattern Analysis and Machine Intelligence (TPAMI)

2022

Learning Two-Step Hybrid Policy for Graph-Based Interpretable Reinforcement Learning
Tongzhou Mu, Kaixiang Lin, Feiyang Niu, Govind Thattai
PDF / Transactions on Machine Learning Research (TMLR)

DialFRED: Dialogue-Enabled Agents for Embodied Instruction Following
Xiaofeng Gao, Qiaozi Gao, Ran Gong, Kaixiang Lin, Govind Thattai, Gaurav Sukhatme
PDF / IEEE Robotics and Automation Letters (RA-L)

Learning to Act with Affordance-Aware Multimodal Neural SLAM
Zhiwei Jia, Kaixiang Lin, Yizhou Zhao, Qiaozi Gao, Govind Thattai, Gaurav Sukhatme
PDF / IROS 2022

2021

LUMINOUS: Indoor Scene Generation for Embodied AI Challenges
Yizhou Zhao, Kaixiang Lin, Zhiwei Jia, Qiaozi Gao, Govind Thattai, Jesse Thomason, Gaurav S.Sukhatme
arXiv / NeurIPS 2021 Workshop on CtrlGen

RCA: A Deep Collaborative Autoencoder Approach for Anomaly Detection
Boyang Liu, Ding Wang, Kaixiang Lin, Pang-Ning Tan, Jiayu Zhou
The 30th International Joint Conference on Artificial Intelligence (IJCAI), 2021
PDF / code

PowerNet: Multi-agent Deep Reinforcement Learning for Scalable Powergrid Control
Dong Chen, Kaian Chen, Zhaojian Li, Tianshu Chu, Rui Yao, Feng Qiu, Kaixiang Lin
IEEE Transactions on Power Systems, 2021
PDF / code

2020

Off-Policy Imitation Learning from Observations
Zhuangdi Zhu, Kaixiang Lin, Bo Dai, Jiayu Zhou
Neural Information Processing Systems (NeurIPS), 2020
PDF / code

Ranking Policy Gradient
Kaixiang Lin, Jiayu Zhou
International Conference on Learning Representations (ICLR), 2020
PDF / code / slides


2018

Efficient Large-Scale Fleet Management via Multi-Agent Deep Reinforcement Learning
Kaixiang Lin, Renyu Zhao Zhe Xu Jiayu Zhou
24th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2018 (Oral)
PDF / code / talk


2017

Privacy-preserving distributed multi-task learning with asynchronous updates
Liyang Xie, Inci M. Baytas, Kaixiang Lin, Jiayu Zhou
23th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2017
PDF


2016

Interactive Multi-Task Relationship Learning
Kaixiang Lin, Jiayu Zhou
The IEEE International Conference on Data Mining series (ICDM), 2016
PDF/ code / slides

Multi-Task Feature Interaction Learning
Kaixiang Lin, Jianpeng Xu, Inci M. Baytas, Shuiwang Ji, Jiayu Zhou
22th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2016
PDF / code

Stochastic convex sparse principal component analysis
Inci M. Baytas, Kaixiang Lin, Fei Wang, Anil K Jain, Jiayu Zhou
EURASIP Journal on Bioinformatics and Systems Biology , 2016
PDF




Template.