Xiao Lin (林啸)

Education

University of Illinois Urbana-Champaign

Ph.D. in Computer Science · Aug. 2023 ~ Jun. 2027 (expected)

Department of Computer Science · Advisor: Prof. Hanghang Tong

Peking University

B.E. in Computer Science · Sep. 2019 ~ Jun. 2023

School of Intelligence Science and Technology · Advisor: Prof. Yisen Wang

Work Experience

Meta

May ~ Aug 2026 · Menlo Park, CA  |  May ~ Oct 2025 · Sunnyvale, CA

Meta Recommendation System · May ~ Aug 2026
Research Intern · Agentic Reinforcement Learning for Deep Search

Core Machine Learning · May ~ Oct 2025
Research Intern · Mixture of Experts for Long-Sequence Recommendation

SenseTime

May ~ Nov 2021 · Beijing, China

Mobile Intelligence Business Team
Research Intern · Architecture Development on Computer Vision

News

Publications

* equal contribution   ^ mentorship

• Conference
• Mixture of Sequence: Theme-Aware Mixture-of-Experts for Long-Sequence Recommendation
  WWW 2026 (Oral)
  Xiao Lin, Robert Tang, Weilin Cong, Mengyue Hang, Kuan Wang, Yajuan Wang, Zhichen Zeng, Ting-Wei Li, Hyokun Yoon, et al.
  Key Insight: Theme-aware MoE captures multi-interest patterns in long user histories.
  Paper Code
  Sequence ModelingRecommender SystemsEfficient AI
• MORALISE: A Structured Benchmark for Moral Alignment in Visual Language Models
  ICML 2026
  Xiao Lin, Zhining Liu, Ze Yang, Gaotang Li, Ruizhong Qiu, Shuhang Wang, Hui Liu, Haotian Li, Sanjay Keswani, et al.
  Key Insight: Structured benchmark probing moral alignment in vision-language models.
  Paper Huggingface
  VLMsMultimodal ReasoningTrustworthy AISafety & AlignmentBenchmarks & Evaluation
• ALERT: Zero-shot LLM Jailbreak Detection via Internal Discrepancy Amplification
  ACL 2026
  Xiao Lin, Philip Li, Zhichen Zeng, Tianyi Li, Tianxin Wei, Xuying Ning, Gaotang Li, Yujun Chen, Hanghang Tong
  Key Insight: Zero-shot jailbreak detection by amplifying internal LLM activation discrepancies.
  Paper
  LLMsAI AgentsTrustworthy AISafety & Alignment
• Mem-Gallery: Benchmarking Multimodal Long-Term Conversational Memory for MLLM Agents
  ACL 2026
  Yuanchen Bei, Tianxin Wei, Xuying Ning, Yanjun Zhao, Zhining Liu, Xiao Lin, Yada Zhu, Hendrik Hamann, Jingrui He, Hanghang Tong
  Key Insight: Benchmark exposing MLLM agents' failure on long-term multimodal memory.
  Paper Code Huggingface
  LLMsVLMsAI AgentsMultimodal ReasoningBenchmarks & Evaluation
• Language in the Flow of Time: Time-Series-Paired Texts Weaved into a Unified Temporal Narrative
  ICLR 2026
  Zihao Li, Xiao Lin, Zhining Liu, Jiaru Zou, Ziyang Wu, Lecheng Zheng, Dongqi Fu, Yada Zhu, Hendrik Hamann, Hanghang Tong, et al.
  Key Insight: Pairing texts and time series into a unified temporal narrative.
  Paper Code
  Multimodal ReasoningSequence Modeling
• Harnessing Consistency for Robust Test-Time LLM Ensemble
  EACL 2026
  Zhichen Zeng, Qi Yu, Xiao Lin, Ruizhong Qiu, Xuying Ning, Tianxin Wei, Yuchen Yan, Jingrui He, Hanghang Tong
  Key Insight: Consistency-driven test-time LLM ensemble for robust multi-model agreement.
  Paper Code
  LLMsTrustworthy AIAdaptation & TTA
• ClimateBench-M: A Multi-Modal Climate Data Benchmark with a Simple Generative Method
  CIKM 2025
  Dongqi Fu, Yada Zhu, Zhining Liu, Lecheng Zheng, Xiao Lin, Zihao Li, Liri Fang, Katherine Tieu, Onkar Bhardwaj, et al.
  Key Insight: Multimodal climate benchmark paired with a simple generative baseline.
  Paper Code
  VLMsMultimodal ReasoningBenchmarks & EvaluationAI for Science
• Breaking Silos: Adaptive Model Fusion Unlocks Better Time Series Forecasting
  ICML 2025
  Zhining Liu, Ze Yang, Xiao Lin, Ruizhong Qiu, Tianxin Wei, Yada Zhu, Hendrik Hamann, Jingrui He, Hanghang Tong
  Key Insight: Adaptive fusion of heterogeneous time series forecasters breaks model silos.
  Paper Code
  Sequence ModelingAdaptation & TTA
• TUCKET: A Tensor Time Series Data Structure for Efficient and Accurate Factor Analysis over Time Ranges
  VLDB 2025
  Ruizhong Qiu, Jun-Gi Jang, Xiao Lin, Lihui Liu, Hanghang Tong
  Key Insight: Tensor data structure enabling efficient factor analysis over time ranges.
  Paper
  Sequence Modeling
• BackTime: Backdoor Attacks on Multivariate Time Series Forecasting
  NeurIPS 2024 Spotlight (Top 2%)
  Xiao Lin, Zhining Liu, Dongqi Fu, Ruizhong Qiu, Hanghang Tong
  Key Insight: First backdoor attack tailored to multivariate time series forecasting.
  Paper Code
  Sequence ModelingGraph LearningTrustworthy AI
• BeMap: Balanced Message Passing for Fair Graph Neural Network
  Learning on Graphs Conference 2024
  Xiao Lin, Jian Kang, Weilin Cong, Hanghang Tong
  Key Insight: Balanced message passing for fairness in graph neural networks.
  Paper
  Graph LearningTrustworthy AI
• Journal
• Flow Matching Meets Biology and Life Science: A Survey
  npj Artificial Intelligence 2026
  Zihao Li*, Zhichen Zeng*, Xiao Lin*, Feihao Fang, Yanru Qu, Zhe Xu, Zhining Liu, Xuying Ning, Tianxin Wei, Ge Liu, Hanghang Tong, Jingrui He
  Key Insight: Survey connecting flow matching to molecular and biological modeling.
  Paper Project
  SurveysAI for Science
• Preprint / Under Review
• Code as Agent Harness
  Preprint
  Xuying Ning, Katherine Tieu, Dongqi Fu, Tianxin Wei, Zihao Li, Yuanchen Bei, Jiaru Zou, Mengting Ai, Zhining Liu, Ting-Wei Li, Lingjie Chen, Yanjun Zhao, Ke Yang, Bingxuan Li, Cheng Qian, Gaotang Li, Xiao Lin, Zhichen Zeng, Ruizhong Qiu, Sirui Chen, Yifan Sun, Xiyuan Yang, Ruida Wang, Rui Pan, Chenyuan Yang, Dylan Zhang, Liri Fang, Zikun Cui, Yang Cao, Pan Chen, Dorothy Sun, Ren Chen, Mahesh Srinivasan, Nipun Mathur, Yinglong Xia, Hong Li, Hong Yan, Pan Lu, Lingming Zhang, Tong Zhang, Hanghang Tong, Jingrui He
  Key Insight: Code as the unifying operational substrate for agent reasoning, action, and execution.
  Paper Project
  LLMsAI AgentsSurveys
• Agentic Reasoning for Large Language Models: A Survey
  Preprint
  Tianxin Wei, Ting-Wei Li, Zhining Liu, Xuying Ning, Ze Yang, Jiaru Zou, Zhichen Zeng, Ruizhong Qiu, Xiao Lin, Dongqi Fu, Zihao Li, Mengting Ai, Duo Zhou, Wenxuan Bao, Yunzhe Li, Gaotang Li, Cheng Qian, Yu Wang, Xiangru Tang, Yin Xiao, Liri Fang, Hui Liu, Xianfeng Tang, Yuji Zhang, Chi Wang, Jiaxuan You, Heng Ji, Hanghang Tong, Jingrui He
  Key Insight: Unified taxonomy of how LLM agents plan, use tools, and reflect to reason.
  Paper Project Huggingface
  LLMsAI AgentsSurveys
• Do VLMs Have a Moral Backbone? A Study on the Fragile Morality of Vision-Language Models
  Preprint
  Zhining Liu, Tianyi Wang, Xiao Lin, Penghao Ouyang, Gaotang Li, Ze Yang, et al.
  Key Insight: VLM moral judgments collapse under simple adversarial perturbations.
  Paper
  VLMsMultimodal ReasoningTrustworthy AISafety & AlignmentBenchmarks & Evaluation
• Subspace Alignment for Vision-Language Model Test-time Adaptation
  Preprint
  Zhichen Zeng, Wenxuan Bao, Xiao Lin, Ruizhong Qiu, Tianxin Wei, Xuying Ning, Yuchen Yan, Chen Luo, Monica Xiao Cheng, et al.
  Key Insight: Subspace-aligned test-time adaptation for vision-language models.
  Paper
  VLMsMultimodal ReasoningTrustworthy AIAdaptation & TTA
• Geometric-Disentanglement Unlearning
  Preprint
  Duo Zhou, Yuji Zhang, Tianxin Wei, Ruizhong Qiu, Ke Yang, Xiao Lin, Cheng Qian, Jingrui He, Hanghang Tong, Heng Ji, Huan Zhang
  Key Insight: Geometric disentanglement enables selective unlearning without collateral damage.
  Paper
  LLMsTrustworthy AISafety & Alignment
• TSAQA: Time Series Analysis Question And Answering Benchmark
  Preprint
  Baoyu Jing, Sanhorn Chen, Lecheng Zheng, Boyu Liu, Zihao Li, Jiaru Zou, Tianxin Wei, Zhining Liu, Zhichen Zeng, Ruizhong Qiu, Xiao Lin, Yuchen Yan, Qi Yu, Dongqi Fu, Jingrui He, Hanghang Tong
  Key Insight: Benchmark evaluating LLM reasoning over real-world time series.
  Paper Code
  LLMsSequence ModelingBenchmarks & Evaluation
• MADE: Graph Backdoor Defense with Masked Unlearning
  Preprint
  Xiao Lin, Mingjie Li, Yisen Wang
  Key Insight: Masked unlearning defends GNNs against stealthy backdoor attacks.
  Paper
  Graph LearningTrustworthy AISafety & Alignment
• Hierarchical LoRA MoE for Efficient CTR Model Scaling
  Preprint
  Zhichen Zeng, Mengyue Hang, Xiaolong Liu, Xiaoyi Liu, Xiao Lin, Ruizhong Qiu, Tianxin Wei, Zhining Liu, et al.
  Key Insight: Hierarchical LoRA experts scale CTR models with parameter efficiency.
  Paper
  Recommender SystemsEfficient AI
• Pave Your Own Path: Graph Gradual Domain Adaptation on Fused Gromov-Wasserstein Geodesics
  Preprint
  Zhichen Zeng, Ruizhong Qiu, Wenxuan Bao, Tianxin Wei, Xiao Lin, Yuchen Yan, Tarek F. Abdelzaher, Jiawei Han, Hanghang Tong
  Key Insight: Gradual graph domain adaptation along fused Gromov-Wasserstein geodesics.
  Paper
  Graph LearningAdaptation & TTA
• Cache Mechanism for Agent RAG Systems
  Preprint
  Shuhang Lin, Zhencan Peng, Lingyao Li, Xiao Lin, Xi Zhu, Yongfeng Zhang
  Key Insight: Cache mechanism that accelerates RAG queries for LLM agents.
  Paper
  LLMsAI AgentsEfficient AI
• CATS: Mitigating Correlation Shift for Multivariate Time Series Classification
  Preprint
  Xiao Lin, Zhichen Zeng, Tianxin Wei, Zhining Liu, Hanghang Tong
  Key Insight: Correlation-shift mitigation for robust multivariate time series classification.
  Paper
  Sequence ModelingGraph LearningAdaptation & TTAEfficient AI

Awards & Service