I am a Ph.D. student in Computer Science at Emory University, advised by Prof. Wei Jin. Previously, I was fortunate to work with Prof. Zhiguang Cao and Dr. Yining Ma. I received my M.S. from the National University of Singapore and completed my undergraduate studies at Nankai University.

My current research focuses on machine learning for time series data. I spend time thinking about:

1. whether there exists a universal architecture for time series, or whether domain-specific inductive biases are necessary;
2. what minimal inductive biases make vanilla Transformers work well on time series;
3. what are the right scaling recipes for time series models (model size, data, context length, or something else);
4. what knowledge transfers across temporal domains, and why large-scale pretraining helps on unseen data;
5. how specialized numerical time series models should collaborate with general-purpose LLM agents.

Previously, I worked on neural combinatorial optimization, where we explored how AI can help solve NP-hard decision-making problems in the real world.

Publications

1. AGDN: Learning to Solve Traveling Salesman Problem with Anisotropic Graph Diffusion Network, Bolin Shen, Ziwei Huang, Zhiguang Cao, and Yushun Dong. ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (SIGKDD), 2026. [Paper] [Code]
2. RADAR: Learning to Route with Asymmetry-aware DistAnce Representations, Hang Yi, Ziwei Huang, Yining Ma, and Zhiguang Cao. International Conference on Learning Representations (ICLR), 2026. [Paper] [Code]
3. PEOAT: Personalization-Guided Evolutionary Question Assembly for One-Shot Adaptive Testing, Xiaoshan Yu, Ziwei Huang, Shangshang Yang, Ziwen Wang, Haiping Ma, and Xingyi Zhang. AAAI Conference on Artificial Intelligence (AAAI), 2026. [Paper]
4. Rethinking Light Decoder-based Solvers for Vehicle Routing Problems, Ziwei Huang, Jianan Zhou, Zhiguang Cao, and Yixin Xu. International Conference on Learning Representations (ICLR), 2025. [Paper] [Code]

News

• [2026.08] Started my Ph.D. in CS at Emory University, advised by Prof. Wei Jin.
• [2026.05] Our paper “AGDN: Learning to Solve Traveling Salesman Problem with Anisotropic Graph Diffusion Network” was accepted by KDD 2026.
• [2026.01] Our paper “RADAR: Learning to Route with Asymmetry-aware DistAnce Representations” was accepted by ICLR 2026.
• [2025.12] Our paper “PEOAT: Personalization-Guided Evolutionary Question Assembly for One-Shot Adaptive Testing” was accepted by AAAI 2026.
• [2025.01] Our paper “Rethinking Light Decoder-based Solvers for Vehicle Routing Problems” was accepted by ICLR 2025.