Research Interests

• Agentic AI safety — long-horizon agent evaluation, behavioral drift, multi-agent systems
• Adversarial ML — jailbreaks, adversarial robustness, risk-aware robust optimization
• Interpretability & reliability — GNN explainability, runtime failure detection, input repair

Other Expertise

• LLM post-training — supervised fine-tuning and reinforcement learning
• AI systems — model development, deployment monitoring, and large-scale data pipelines (Spark, ClickHouse)
• Distributed training — multi-GPU training and distributed reinforcement learning with Ray

Publications

See also Google Scholar.

• [P5]
  SWE-Continuum: Multi-Stage Profiling of Behavioral Continuity in Long-Horizon Agents

  Under Review at ICLR · 2027

  Changsheng Sun, Ruihan Yang, Yijun Zeng, Hao Cui, Haochun Liao, Hanna Seah, Linxi Yu, Xiaosong Gao, Xinke Li

  pdf
• [C6]
  From Attack Surfaces to Actual Operations: A Survey of Modern LLM Jailbreaks

  ACL Findings · 2026

  Ruikang Zhou, Changsheng Sun, Mark Huasong Meng

  pdf
• [P4]
  Risk-Aware Robust Graph Network Explanation

  Under Review · 2026

  Changsheng Sun, Xinke Li, Jin Song Dong

  pdf
• [C5]
  Ignoring Directionality Leads to Compromised Graph Neural Network Explanations

  IEEE S&P Workshops (HMI-SA) · 2025

  Changsheng Sun, Xinke Li, Jin Song Dong

  pdf
• [C4]
  PointCVaR: Risk-optimized Outlier Removal for Robust 3D Point Cloud Classification

  AAAI (Oral) · 2024

  Xinke Li, Junchi Lu, Henghui Ding, Changsheng Sun, Joey Tianyi Zhou, Chee Yeow Meng

  pdf code
• [C3]
  Repairing Failure-inducing Inputs with Input Reflection

  ASE · 2022

  Yan Xiao, Yun Lin, Ivan Beschastnikh, Changsheng Sun, David S. Rosenblum, Jin Song Dong

  pdf
• [J1]
  PIANO: Influence Maximization Meets Deep Reinforcement Learning

  IEEE TCSS · 2022

  Hui Li, Mengting Xu, Sourav S. Bhowmick, Shafiq Joty Rayhan, Changsheng Sun, Jiangtao Cui

  pdf
• [C2]
  Self-Checking Deep Neural Networks in Deployment

  ICSE · 2021

  Yan Xiao, Ivan Beschastnikh, David S. Rosenblum, Changsheng Sun, Sebastian Elbaum, Yun Lin, Jin Song Dong

  pdf code
• [P3]
  Generalizing Neural Networks by Reflecting Deviating Data in Production

  arXiv · 2021

  Yan Xiao, Yun Lin, Ivan Beschastnikh, Changsheng Sun, David S. Rosenblum, Jin Song Dong

  pdf
• [C1]
  Digraph Inception Convolutional Networks

  NeurIPS · 2020

  Zekun Tong, Yuxuan Liang, Changsheng Sun, Xinke Li, David S. Rosenblum, Andrew Lim

  pdf code
• [P2]
  Directed Graph Convolutional Network

  arXiv · 2020

  Zekun Tong, Yuxuan Liang, Changsheng Sun, David S. Rosenblum, Andrew Lim

  pdf
• [P1]
  Disco: Influence Maximization Meets Network Embedding and Deep Learning

  arXiv · 2019

  Hui Li, Mengting Xu, Sourav S. Bhowmick, Changsheng Sun, Zhongyuan Jiang, Jiangtao Cui

  pdf