cv
Education, research experience, and publications.
Basics
| Name | Wenyu Huang |
| Label | PhD Candidate in Natural Language Processing |
| wyhuang1998@gmail.com | |
| Url | https://hwy9855.github.io/ |
| Summary | Final-year PhD candidate at the University of Edinburgh researching retrieval-augmented generation, model memory, multi-hop reasoning, and language agents. |
Work
-
2026.04 - 2026.06 Cambridge, UK
Research Intern, Machine Intelligence
Microsoft Research Cambridge
Worked with John Winn on memory-augmented language modeling and the parameter efficiency of external factual memory.
-
2025.07 - 2026.03 United Kingdom
Research Intern
Huawei Technologies Research & Development (UK) Ltd
Researched reinforcement learning for language-model agents.
-
2022.06 - 2022.11 United Kingdom
Research Intern
Huawei Technologies Research & Development (UK) Ltd
Developed methods for large-scale product-offer alignment; the resulting work was published at SIGIR 2024.
Education
Publications
-
2026.04.24 Memory-T1: Reinforcement Learning for Temporal Reasoning in Multi-session Agents
ICLR 2026, Poster
Temporal reasoning and memory in multi-session language agents.
-
2025.07.01 Rethinking Stateful Tool Use in Multi-Turn Dialogues: Benchmarks and Challenges
Findings of ACL 2025
Introduced benchmarks for evaluating stateful tool use over long-horizon dialogues.
-
2025.07.01 Masking in Multi-hop QA: An Analysis of How Language Models Perform with Context Permutation
ACL 2025 (Main Conference), Oral
Oral presentation; context ordering, causal masking, and bidirectional attention for multi-hop reasoning.
-
2025.05.15 Prompting Large Language Models with Knowledge Graphs for Question Answering Involving Long-tail Facts
Knowledge-Based Systems, 2025
Introduced LTGen and evaluated knowledge-graph and passage evidence for long-tail factual question answering.
-
2025.05.01 Rethinking Memory in AI: Taxonomy, Operations, Topics, and Future Directions
arXiv, 2025
Equal contribution; a taxonomy of memory representations, operations, and research directions.
-
2024.11.01 Less is More: Making Smaller Language Models Competent Subgraph Retrievers for Multi-hop KGQA
Findings of EMNLP 2024
Introduced efficient generative subgraph retrieval for knowledge-graph RAG.
-
2024.07.01 A Large-scale Offer Alignment Model for Partitioning, Filtering and Matching Product Offers
SIGIR 2024
A scalable pipeline for partitioning, filtering, and matching product offers.
-
2024.01.01 UniMS-RAG: A Unified Multi-source Retrieval-Augmented Generation for Personalized Dialogue Systems
arXiv, 2024
Unified retrieval across heterogeneous information sources for personalized dialogue.
-
2023.11.01 Retrieval Augmented Generation with Rich Answer Encoding
IJCNLP-AACL 2023
Rich answer representations for retrieval-augmented generation in knowledge-intensive NLP.