University of Sheffield  ·  School of Computer Science

Overview

Our interactive NLP research lab advances the frontiers of natural language processing by architecting intelligent systems that actively learn, reason, and evolve through multi-dimensional interactions, driving toward a future where language technologies seamlessly support every aspect of human life.

We pursue this mission through principled research across four interaction pillars:

 Environment

Grounding language processing in diverse real-world scenarios, spanning open-ended digital ecosystems, interactive simulations, and dynamic workflows, where intelligent systems must actively perceive context, adapt, and execute actions.

 Human

Building collaborative systems that deeply align with user intent, learn from real-time feedback, and fluidly co-evolve with users, positioning human needs at the core of AI to intuitively support every aspect of daily life.

 Knowledge

Integrating massive, heterogeneous information sources, from web-scale documents and structured knowledge bases to complex ontologies, powering language models with robust knowledge-grounded generation and verifiable, fact-grounded reasoning.

 Agents

Engineering autonomous multi-agent networks where specialized language models communicate, collaborate, and critique one another—unlocking collective intelligence to solve highly complex, large-scale challenges.

Our work spans Conversational AI, Information Retrieval, Personalisation, and AI for Mental Health, and has been published at top venues including ACL, EMNLP, NAACL, ICLR, SIGIR, AAAI, CIKM and ECIR.

Research Themes

  • Proactive Conversational Information Seeking — systems that ask targeted questions to resolve uncertainty and satisfy complex information needs
  • Deep Research / Retrieval-Augmented Generation (RAG) — Trustworthy generations that ground LLM outputs in retrieved evidence and/or iterative reasoning.
  • User Simulation — synthetic users for scalable evaluation of dialogue and IR systems
  • Conversational Recommendation — preference elicitation and cold-start personalisation through dialogue
  • AI for Mental Health — LLM-based tools for depression screening and emotionally-aware conversation

Team

Current PhD Students

Feng Xia Proactive Conversational Recommendation

Affiliated PhD Students (Co-Supervised)

Xingyu Deng Scientific Fact Checking co-supervised with Dr Mark Stevenson
Mingzi Cao Model Expansion for Large Language Models co-supervised with Prof. Nikos Aletras
Yanwen Peng Communication in Multi-Agent Systems co-supervised with Prof. Nikos Aletras

Postgraduate (MSc/MRes) Students

Vismay Cheelaganahalli Vijaya Kumar Conversational Asking Clarification Questions
Virthi Murali Unsupervised Adaptive RAG

Undergraduate & Project Students

Yacine Zadi Conversational Movie Recommendation
Benjamin Foggo Asking Clarification Questions
Dylan Harding-Murphy Adaptive RAG
Callum Pringle Proactive Conversational Actions

Join Us

We are actively recruiting PhD students for 2026/27 entry. If you are interested in working on:

  • Proactive conversational information seeking
  • Deep Research & Retrieval-Augmented Generation
  • User simulation for IR and dialogue
  • Automatic conversational support for mental health

please email xi.wang@sheffield.ac.uk with the subject line [PhD 26/27].

For undergraduate and postgraduate project opportunities, please get in touch to discuss available topics.