Yaxiong Wu
Yaxiong Wu received his PhD from the School of Computing Science at the University of Glasgow (UK), under the joint supervision of Prof. Craig Macdonald and Prof. Iadh Ounis. His doctoral research focused on multi-modal conversational recommendation, with an emphasis on closed-loop systems and reinforcement learning. During his PhD, he was fully funded by the EPSRC Closed Loop Data Science Project and the Computing Science Excellence Bursary. After completing his doctorate, he spent two years as a Senior Staff Engineer at Noah's Ark Lab, Huawei (Singapore), where he conducted research on AI Agents. His current research interests primarily lie in the following areas:
- Information Retrieval
- Large Language Models
- Agentic AI
Work Experience
2023/10 — 2025/09
Senior Staff Engineer, Search & Rec. Lab, Noah's Ark Lab, Huawei, Singapore.
AI Agents (Agent Memory, Recommender Agent, Agentic RAG)
Education
2019/10 — 2023/09
University of Glasgow (UofG) — Glasgow, UK
Ph.D. in Computing Science, supervised by Prof. Craig Macdonald & Prof. Iadh Ounis.
2018/05 — 2019/09
University of Glasgow (UofG) — Glasgow, UK
M.Sc. in Data Science, Graduation with Distinction.
2013/09 — 2016/03
Beihang University (BUAA) — Beijing, China
M.Eng. in Aerospace Science and Technology, supervised by Prof. Haixing Wang.
2009/08 — 2013/07
Harbin Engineering University (HEU) — Harbin, China
B.Eng. in Flight Vehicle Propulsion Engineering, B.B.M. in Business Administration (minor).
Publications (Selected)
-
2026Thought-Augmented Planning for LLM-Powered Interactive Recommender Agent Haocheng Yu, Yaxiong Wu, Hao Wang, Wei Guo, Yong Liu, Yawen Li, Yuyang Ye, Junping Du, Enhong Chen. KDD 2026 (Accepted). 📄 [Paper]
-
2025Query-Centric Graph Retrieval Augmented Generation Yaxiong Wu, Jianyuan Bo, Yongyue Zhang, Sheng Liang, Yong Liu. ArXiv. 📄 [Paper]
-
2025SGMem: Sentence Graph Memory for Long-Term Conversational Agents Yaxiong Wu, Yongyue Zhang, Sheng Liang, Yong Liu. ArXiv. 📄 [Paper]
-
2025Effective and Efficient Schema-aware Information Extraction Using On-Device Large Language Models Zhihao Wen, Sheng Liang, Yaxiong Wu, Yongyue Zhang, Yong Liu. ArXiv. 📄 [Paper]
-
2025Adaptive Schema-aware Event Extraction with Retrieval-Augmented Generation Sheng Liang, Hang Lv, Zhihao Wen, Yaxiong Wu, Yongyue Zhang, Hao Wang, Yong Liu. EMNLP 2025 Findings. 📄 [Paper]
-
2025From Human Memory to AI Memory: A Survey on Memory Mechanisms in the Era of LLMs Yaxiong Wu, Sheng Liang, Chen Zhang, Yichao Wang, Yongyue Zhang, Huifeng Guo, Ruiming Tang, Yong Liu. ArXiv. 📄 [Paper]
-
2024Personalised Multi-Modal Interactive Recommendation with Hierarchical State Representations Yaxiong Wu, Craig Macdonald, Iadh Ounis. ACM TORS. 📄 [Paper] 🛠️ [Code] 📂 [Dataset] 👤 [User Simulator]
-
2022Multi-Modal Dialog State Tracking for Interactive Fashion Recommendation Yaxiong Wu, Craig Macdonald, Iadh Ounis. ACM RecSys 2022. 📄 [Paper]
-
2022Multimodal Conversational Fashion Recommendation with Positive and Negative Natural-Language Feedback Yaxiong Wu, Craig Macdonald, Iadh Ounis. ACM CUI 2022. 📄 [Paper]
-
2021Partially Observable Reinforcement Learning for Dialog-based Interactive Recommendation Yaxiong Wu, Craig Macdonald, Iadh Ounis. ACM RecSys 2021. 📄 [Paper]
-
2020A Hybrid Conditional Variational Autoencoder Model for Personalised Top-N Recommendation Yaxiong Wu, Craig Macdonald, Iadh Ounis. ACM ICTIR 2020. 📄 [Paper]
Patents
-
2025/07CoSteer: Collaborative Decoding-Time Personalization via Local Delta Steering Co-Inventor (2nd). Patent ID: 92093217CN01.
-
2025/05Query-Centric Graph Retrieval Augmented Generation Lead Inventor (1st). Patent ID: 92087577CN01.
-
2024/10Thought-Augmented Planning for LLM-Powered Interactive Recommender Agent Lead Inventor (1st). Patent ID: 92074271CN01. (KDD 2026)
-
2024/09Adaptive Schema-aware Event Extraction with Retrieval-Augmented Generation Lead Inventor (1st). Patent ID: 92073058CN01. (EMNLP 2025 Findings)
-
2024/07Hierarchical Skill-based Collaborative Multi-Agent System for Interactive Recommendation Lead Inventor (1st). Patent ID: 92069447CN01.
-
2024/05FuxiAgent: A Lightweight Multi-Agent Framework Co-Inventor (2nd). Patent ID: 92064791CN01.
Honors & Scholarships
-
2025Excellent Individual Award 2024 (Singapore Research Center, Huawei)
-
2024Project Technology Innovation Award (Singapore Search & Reccommendation Lab, Huawei)
-
2019 — 2023Closed-Loop Data Science Scholarship (CoSE, UofG; EPSRC Funded)
-
2019 — 2022Computing Science Excellence Bursary (CoSE, UofG)
-
2019Graduation with Distinction (CoSE, UofG)
-
2016Excellent Master's Degree Dissertation (SoA, BUAA)