Dr Diego Perez-Liebana

Research

Artificial Intelligence, Games, Tree Search, Evolutionary Computation

Interests

My research is centred in the application of Artificial Intelligence to games, Reinforcement Learning and Evolutionary Computation. At the moment, I am especially interested on General Video Game Playing, which involves the creation of content and agents that play any real-time game that is given to it. He has published more than 60 papers in leading conferences and journals with 2 best paper awards, a BCS Intelligence Prize, and 1,900+ citations. Applying search algorithms, evolutionary computation and reinforcement learning to games, he is the main organizer of popular game AI competitions (GVGAI, MARLO) - securing substantial industry funding by MSR and Google Deepmind, general chair of the 2019 IEEE Conference on Games, track chair of academic (IEEE-CEEC, IEEE-CIG, ACM-FDG) and industry (nucl.ai) game conferences, organiser of game AI workshops and meetups, and frequently invited speaker at industry events like the London Games Festival. He is currently co-supervising 5 PhD students. As QMUL?s IGGI industry liaison, he is drawing on his extensive games industry experience, with several published PC and console titles and developed game AI tools under his belt.

Publications

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

Grants

Research Group

PhD Students

• Yu-Jhen Hsu
  Developing Agents For Digital Games Has Been Receiving Increasing Attention From Academia With The Advance of Artificial Intelligence (AI). in Such Games, An Agent Is Expected to Perform Actions and Fulfil Its Goals. as The Strategy Game Is Complex in Terms of The Game State Space and Action Space, Implementing Intelligent Agents Has Been Regarded as a Challenging Topic. Search Algorithms, E.G. Monte Carlo Tree Search (McTs), Are Commonly Used in Existing Works to Design Agents. However, Practical Issues, Such as The High Game Complexity, Hinder Its Application and Introduce Open Research Problems. The Proposed Study Will Focus on Improving McTs. We Investigate How to Determine The Convergence of The McTs, How to Measure Agents’ Decisions and How to Build Its Awareness to Realize Authorial Control. on Successful Completion, The Research Will Have a Great Impact on Both The Academia and Game Industry, Motivating Further Research Directions.
• Dominik Jeurissen
  Large Language Models (Llms) Have Sparked a Remarkable Revolution in AI, Demonstrating Reasoning Abilities That Exceeded The Expectations of Many Researchers. Research on Games Like Minecraft Suggests That Llms Knowledge About Games Allows Them to Play Games Without Any Pre-Training Required. However, Most Research on Llm-Powered Game-Playing Agents Has Utilized Them Primarily For Planning. However, Complex Strategy Games Like Total War Require a Lot of Dynamic Decision-Making, Requiring The Agent to Constantly Adapt to Their Current Situation. in This Study, We Aim to Address This Gap by Exploring The Potential of Llms For Playing Complex Strategy Games and How Llms Could Be Used to Playtest These Games. We Will Review The Current State of Research on Llm-Powered Agents, Identify Their Limitations, and Present Our Project Proposal. Finally, We Will Present Our Plans For Achieving This Goal.
• Nathan John-McDougall
  Adversarial Behaviour Debugging With Neuroevolution
• Joshua Kritz
  The Use of AI Strategy Data in Redefining Modern Tabletop Game Design
• George Long
  Optimisation in Role-Playing and Tabletop Games Games
• Juexi Shao
  Grounding of Conversational Agents in a Collaborative Construction Environment
• Sunny Thaicharoen
  Quantification of Moba Metagames to Support Gameplay Balancing and Design
• Marko Tot
  Search, Learning, and World Models: Strengths, Limitations, and Ways Forward
• Linjie Xu
  Efficient Data Usage in Planning and Reinforcement Learning

News