Dr. Haiyan Yin

Team Lead of Agentic AI

A*STAR Centre for Frontier AI Research (CFAR), Singapore

Research Areas: LLM Agents & Autonomous Workflows • Reinforcement Learning • Trustworthy AI & Verifiable Control • Continual & Meta-Learning
yinhaiyan@outlook.com / yin_haiyan@a-star.edu.sg | Google Scholar | LinkedIn
Haiyan Yin

About Me

I am a Senior Scientist and Team Lead of Agentic AI at the Centre for Frontier AI Research (CFAR), A*STAR, working with Prof. Ivor Tsang. My research bridges sequential decision-making and language model reasoning to develop scalable, verifiable, and self-evolving agents capable of long-horizon reasoning, autonomous adaptation, and reliable execution in the open world.

Before joining A*STAR, I was a Research Scientist at Sea AI Lab (SAIL) and previously worked at Baidu Research USA in Seattle. I obtained both my Ph.D. and B.Eng. degrees from the College of Computing and Data Science (CCDS), Nanyang Technological University (NTU) where I was fortunate to be advised by Prof. Sinno Jialin Pan.

Recent Activity

[May 2026]

Delivered a Keynote Speech at the VALSE 2026 Workshop on Neural Symbolic AI on "From Reactive Reasoning to Verifiable Control: Pathway to Neuro-Symbolic Autonomous Agents".

[May 2026]

Recognized as an 🏆 ICML 2026 Gold Reviewer.

[May 2026]

Two papers on causally-inspired physical reasoning in VLMs and skill refinement for long-horizon web agents have been accepted to KDD 2026.

[Apr 2026]

Two papers on neural-symbolic skill induction for long-horizon planning and evolutionary counterfactual reasoning for multi-agent planning have been accepted for presentation at ICML 2026 (Seoul, South Korea). I will be attending the conference in person.

[Jan 2026]

Four papers have been accepted to ICLR 2026.

[Jan 2026]

Serving as the Leading Organizer for the AAAI 2026 Workshop on Trust and Control of Agentic AI .

[Jan 2026]

Received the A*STAR CDF Best Presentation Award (with SGD $1,500 travel grant).

Openings

Active Recruitment & Supervision

[📢Opportunities] Research Interns (Remote/On-site), Visiting Scholars, & Postdocs

We are actively looking for highly motivated minds to work on high-impact publications. We prioritize candidates with outstanding programming skills and academic drive.

Selected Working Topics:
• Trustworthy Agentic AI & Verification: Neural-symbolic tracing, logical grounding, agent verification.
• Self-Evolving Systems & Skill Recovery: Memory-guided planning, failure tracing and automated repair.
• Reinforcement Learning: Post trainning & tool calling, embodied policy learning and continual RL.
• AI4Science: Programmatic LLM agents for physics-driven catalysis and complex chemical surface discovery.
Submission Guide: Interested candidates (e.g., visiting scholars, local/overseas master studnets, junior researchers) please email your CV, publication list/drafts, and target research statement directly to yinhaiyan@outlook.com.

Publications

Show All on Google Scholar | (* joint first author, † corresponding author)
ICML 2026

Lifting Traces to Logic: Programmatic Skill Induction with Neuro-Symbolic Learning for Long-Horizon Agentic Tasks

Jie-Jing Shao, Haiyan Yin†, Yueming Lyu, Xingrui Yu, Lan-Zhe Guo, Ivor Tsang, James Kwok, Yu-Feng Li†.

[Paper] [Code Base]
Programmatic Skill Induction • Neuro-Symbolic Control • Agentic Workflow
ICML 2026

EvoCF: Multi-agent Collaboration with Memory-guided Evolutionary Counterfactual Planning

Haotian Chi, Zeyu Feng, Xingrui Yu, Linbo Luo, Yew-Soon Ong, Ivor Tsang, Hechang Chen†, Yi Chang†, Haiyan Yin†.

[Paper]
Multi-Agent Planning • Evolutionary Programming • Counterfactual Reasoning • Agentic Memory
KDD 2026

SkillTracer: Structural Failure Attribution and Refinement of Agentic Skills in Long-Horizon Web Tasks

Yuyang Li*, Yiran Dou*, Jie-Jing Shao, Yueming Lyu, Ivor Tsang, Haiyan Yin†.

[Paper] [Project Page]
Agentic Skill Refinement • Failure Attribution • Long-Horizon Web Agents
KDD 2026

Causal Scaffolding for Physical Reasoning: A Benchmark for Causally-Informed Physical World Understanding in VLMs

Tianyi Tang, Zhuoyi Lin†, Zeyu Feng, Tianyi Ma, Yew-Soon Ong, Ivor Tsang, Haiyan Yin†.

[Paper] [Benchmark]
Physical World Understanding • Causal Reasoning • Vision-Language Models
ICLR 2026

FlowSearcher: Synthesizing Memory-Guided Agentic Workflows for Web Information Seeking

Keyi Xiang, Zeyu Feng, Zhuoyi Lin†, Yueming Lyu, Boyuan Shi, Yew-Soon Ong, Ivor Tsang, Haiyan Yin†.

[Paper]
Agentic Workflows • Deep Research • Multi-level Memory
ICLR 2026

Masked Skill Token Training for Hierarchical Off-Dynamics Transfer

Zeyu Feng, Haiyan Yin, Yew-Soon Ong, Harold Soh.

[Paper]
Hierarchical Reinforcement Learning • Diffusion Models • Off-Dynamics Generalization
ICLR 2026

Verification and Co-Alignment via Heterogeneous Consistency for Preference-Aligned LLM Annotations

Cheng Chen, Haiyan Yin, Ivor Tsang.

[Paper]
Consistency Verification • LLM Preference Alignment
NeurIPS 2025

Instruct Flow: Adaptive Symbolic Constraint-Guided Code Generation for Long-Horizon Planning

Haotian Chi, Zeyu Feng, Yueming Lyu, Chengqi Zheng, Linbo Luo, Yew-Soon Ong, Ivor Tsang, Hechang Chen†, Yi Chang†, Haiyan Yin†.

[Paper] [Code]
Code Generation • Symbolic Constraint Induction • Robotic Manipulation • Embodied Planning
ICML 2025 MAS Workshop

MermaidFlow: Redefining Agentic Workflow Generation via Safety-Constrained Evolutionary Programming

Chengqi Zheng, Jianda Chen, Yueming Lyu, Wen Zheng Terence Ng, Haopeng Zhang, Yew-Soon Ong, Ivor Tsang, Haiyan Yin*.

[Paper] [Code]
Agentic Workflow Optimization • Self-Evolving Agentic Systems • Multi-Agent Systems

Selected Workshop Papers

[ICLR '26 AIWILD]

NesyProAct: Proactive Neural-Symbolic Control for Web Agents

Tianyi Tang, Keyi Xiang, Jie-Jing Shao, Yueming Lyu, Ivor Tsang, Yew-Soon Ong, Haiyan Yin†.

[ICLR '26 VerifAI2]

Grounding Long-Horizon Multi-Agent Planning in GUI Environments with Shared Contract-based DAGs

Hao Yu, Weiming Li, Yueming Lyu, Jie-Jing Shao, Yulei Sui, Ivor Tsang, Haiyan Yin†.

Professional Service & Rewards

Organizing Roles

  • Leading Organizer, AAAI 2026 Workshop on Trust and Control of Agentic AI.
  • Organizing Committee, A*STAR-NVIDIA Joint Workshop on Agentic GenAI (2025).
  • Workflow Chair, IEEE Conference on Artificial Intelligence (CAI) (2024).

Reviewing & Honors

  • 🏆 Gold Reviewer, ICML 2026.
  • 🏆 Highlighted Reviewer, ICLR 2022.
  • 🇸🇬 SG100 Woman in Tech (2025).
  • Reviewer for NeurIPS, ICML, ICLR, KDD, AAAI, IJCAI, TPAMI, TMLR, etc
  • A*STAR CDF Best Presentation Award (2025).

Supervisions & Joint Work

I prioritize rigorous mentorship and strive to cultivate an open research environment where curiosity, intellectual independence, and bold ideas can thrive.

Active Student Co-authors & Intern Placement:
  • • Li Shenggui (Ph.D. Candidate @ NTU) — Scaling & video generation inference.
  • • Jie-Jing Shao (Visiting Scholar @ LAMDA, Nanjing University) — Programmatic skill induction & LLM Agents.
  • • Li Yuyang (Master Intern @ NUS) — Long-horizon agentic control & harness engineering.
  • • Haotian Chi (Ph.D. Intern @ JLU) — Multi-agent counterfactual planning & code generation.
  • • Tang Tianyi (Master Intern @ NTU) — Agentic memory & physics-driven agent control.
  • • Keyi Xiang (Master Intern @ NTU) — Agentic workflow optimization & deep research.