Search Results

< Back James Gardner University of York iGGi PG Researcher I am a third-year PhD student at The University of York, specialising in computer vision and machine learning for 3D scene understanding. Supervised by Dr William Smith, my research focuses on neural-based vision and language priors in inverse rendering and scene representation learning. I'm particularly interested in neural fields, generative models, 3D computer vision, differentiable rendering, geometric deep learning, multi-modal models, and 3D scene understanding in general. My research has been recognised with publications at prestigious conferences including NeurIPS and ECCV. Currently, I am working as a research fellow on the ALL.VP project, funded by BridgeAI and Dock10, developing relightable green screen performance capture using deep learning and inverse rendering techniques. This work aims to provide greater creative control to film and TV productions without requiring expensive LED volumes or post-production. I hold an MEng in Electronic Engineering from The University of York, for which I was awarded the IET Prize for outstanding performance and the Malden Owen Award for the best-graduating student on an MEng programme. A description of James' research: My research lies at the intersection of computer vision, machine learning, and 3D scene understanding, with a particular focus on neural-based approaches and the integration of vision and language priors. My work spans a range of topics including neural fields, generative models, differentiable rendering, and geometric deep learning. A key theme in my research is the use of 3D inductive biases for inverse rendering, addressing challenges such as illumination estimation, albedo/geometry disentanglement, and shadow handling in complex outdoor scenes. I've made contributions in creating a rotation-equivariant neural illumination model and spherical neural models for sky visibility estimation in outdoor inverse rendering. Additionally, my work extends to learning rotation-equivariant latent representations of the world from 360-degree videos, aimed at advancing the field of 3D scene understanding and developing models with an understanding of core physical principles such as object permanence. Through my research, I aim to build computer systems capable of deeply comprehending the 3D world, utilising self-supervised, generative, and non-generative approaches to push the boundaries of what's possible in computer vision and scene representation learning. james.gardner@york.ac.uk Email Mastodon Other links Website https://www.linkedin.com/in/jadgardner/ LinkedIn BlueSky https://jadgardner.github.io/ Github Featured Publication(s): The Sky's the Limit: Relightable Outdoor Scenes via a Sky-Pixel Constrained Illumination Prior and Outside-In Visibility Themes Game AI - Previous Next

< Back Dr Charline Foch University of York iGGi Alum Charline first came to the UK in 2011 to study English and Film Studies at King’s College London, before going on to a MSc in Film, Exhibition and Curation at the University of Edinburgh. By chance, accident or fate, she stumbled into the games industry, working in an independent game studio in Berlin, where she touched upon customer support, community management, content writing and QA for a new MMORPG. This experience gave her the push to start a PhD in video games. In her spare time, she is an avid film viewer, volleyball player, and amateur artist. Charline’s research focuses on how people conceptualise failure, with an emphasis on its perceived positive, desirable effects on player experience. Throughout her PhD, she has conducted research among video games players to gain a better understanding of what they perceive as the purpose and value of failure in the games they play; and conducted research among video games developers to gain a better understanding of what processes, obstacles, and ideas go into the design and implementation of failure in their games. With a focus on single-player, more narrative-driven games, she has used this research to design a cards-based design toolkit to support game designers in approaching the question of fail states and player experience in the early stages of the game development process, helping them reflect on the intersection between failure, game mechanics, storytelling, and player experience when working on their games. Aside from her PhD, Charline has also worked with the Digital Creativity Labs on the PlayOn! project, a European project gathering 9 theatres across Europe working on immersive technologies (VR, AR, apps for audience participation...) and theatre productions. During her time at PlayOn!, she has worked on the connections between the games industry and the performance arts, investigating how technology, game design principles, and theatre can work together, and what barriers practitioners face when attempting to reconcile all sides in a single production through experimentation and collaboration. charline.foch@york.ac.uk Email https://mastodon.gamedev.place/@chafoch Mastodon https://charlinefoch.carrd.co Other links Website https://www.linkedin.com/in/charline-foch-97196663 LinkedIn BlueSky Github Supervisor: Dr Ben Kirman Featured Publication(s): “The game doesn't judge you”: game designers’ perspectives on implementing failure in video games “Slow down and look”: Desirable aspects of failure in video games, from the perspective of players. Themes Design & Development Player Research - Previous Next

< Back Dr Maximilian Croissant University of York iGGi Alum I’m a psychology researcher, writer and game designer, exploring our emotional connection with games and creating games with purpose. Coming from a B.Sc. and M.Sc. in psychology and neuroscience, I’m now at the intersection of emotion research, design, and human-computer interaction and try to build design-oriented solutions for adapting game content to affective data. My project will include theoretical groundwork, investigating the emotional relationship between player and games and from there build an affective fear-focused VR horror game with specific and practical solutions in terms of emotion measurement, modelling, and adaptation. The ultimate goal is to help fill knowledge gaps that currently hold us back on making commercially viable affective games and provide tools to design games for a deep emotional impact. I’m also the Co-Founder of Vanilla Noir, a small studio working on applied games that aim to promote well-being and satisfying user experiences. For me, games are a great tool to explore psychological phenomena through interactions and the design and development of games based on applied psychology has great potential to help make the world a bit of a better place. mc2230@york.ac.uk Email http://www.maximilian-croissant.de/en Mastodon https://www.vanilla-noir.com Other links Website https://www.linkedin.com/in/maximilian-croissant LinkedIn BlueSky https://gitlab.com/MaximilianCroissant Github Supervisor(s) Dr Cade McCall Featured Publication(s): Advancing Methodological Approaches in Affect-Adaptive Video Game Design: Empirical Validation of Emotion-Driven Gameplay Modification Using Virtual Reality to Investigate the Influence of Sleep Deprivation on In-the-Moment Arousal During Exposure to Prolonged Threats Affective Systems: Progressing Emotional Human-Computer Interactivity with Adaptive and Intelligent Game Systems An appraisal-based chain-of-emotion architecture for affective language model game agents Emotion Design for Video Games: A Framework for Affective Interactivity Theories, methodologies, and effects of affect-adaptive games: A systematic review A data-driven approach for examining the demand for relaxation games on Steam during the COVID-19 pandemic Endocannabinoid concentrations in hair and mental health of unaccompanied refugee minors Progress in Adaptive Web Surveys: Comparing Three Standard Strategies and Selecting the Best Themes Design & Development Player Research - Previous Next

< Back Dr Henrik Siljebråt Goldsmiths iGGi Alum Henrik has a background in IT/DevOps and a Masters in Cognitive Science from Lund University. Based on established neurobiological correlates of reinforcement learning (RL), I investigate animal learning and decision making using cognitive modeling techniques, such as probabilistic programming and machine learning. Animals somehow manage to create useful representations of incoming sensory information, representations then used for learning and decision making. How these representations of states of the world are integrated into task structure and models of the world is an open question, which I investigate using behavioural experiments with humans and bumblebees and modelling said behaviour using RL combined with hidden state models for representing states and task structure. The potential findings of these experiments have promise to not only elucidate the workings of the animal brain but also provide valuable contributions to artificial intelligence, where improved models of state representations could vastly improve data efficiency and generalizability over current generation systems. Please note: Updating of profile text in progress h.siljebrat@gold.ac.uk Email Mastodon https://henrik.siljebrat.se Other links Website https://www.linkedin.com/in/henrik-siljebrat LinkedIn BlueSky https://github.com/fohria Github Featured Publication(s): On State Representations and Behavioural Modelling Methods in Reinforcement Learning The Effect of State Representations in Sequential Sensory Prediction: Introducing the Shape Sequence Task Towards human-like artificial intelligence using StarCraft 2 Themes - Previous Next

< Back Dr Piers Williams University of Essex iGGi Alum Partial Observability as a game mechanic There is a wide variety of different types of games, each providing its own unique challenge to artificial intelligence. Not all games provide full access to the environment, creating interest and difficulty by hiding particular pieces of information from the player. Other types of game expect teamwork from the players rather than being solely adversarial. Some games use both restrictions, and it is this type of game that this thesis concentrates on. Piers graduated from the University of Essex with an MSc in Computer Science. His interests lie in the field of Artificial Intelligence and in particular Multi-Agent Systems. Please note: Updating of profile text in progress Email Mastodon Other links Website LinkedIn BlueSky Github Featured Publication(s): Hexboard: A generic game framework for turn-based strategy games Evaluating and Modelling Hanabi-Playing Agents Monte carlo tree search applied to co-operative problems The 2018 hanabi competition Artificial intelligence in co-operative games with partial observability Ms. Pac-Man Versus Ghost Team CIG 2016 Competition Cooperative games with partial observability Themes Game AI - Previous Next

< Back Prof. Simon Colton Queen Mary University of London iGGi Co-Investigator Supervisor Simon Colton is an AI researcher with particular focus on issues of Computational Creativity, where we engineer software to take on creative responsibilities in art and science projects. He undertakes projects advancing the state of the art in generative technologies such as evolutionary approaches and deep learning, and uses these to help develop software such as The Painting Fool, The WhatIf Machine, the Wevva game designer, the HR3 automated code generator, and the Art Done Quick casual creator for visual art. In turn, these software systems and their output are used in cultural projects such as a poetry readings, art exhibitions, game jams, and even the production of a West-End musical. This enables Simon to undertake much public engagement, with coverage from the BBC, The Guardian, MIT Tech Review, The New Scientist and many others. These practical and cultural projects inform an evolving philosophical discourse around what it means for machines to be creative, and Simon has co-authored numerous essays driving forward our understanding of this important topic. In this way, he has helped to introduce ideas such as automated framing of products and processes, issues of authenticity and the notion of the machine condition, i.e., what the lived experience of a machine is, and how this could be expressed by that machine through creative production. He is particularly interested in supervising students in project where we apply generative technologies to applications in videogame design, visual art, software engineering, music and text generation. One particular current interest is stretching the boundaries of both what can be achieved by, and our understanding of, generation deep learning methods such as generative adversarial networks (GANs) and auto encoders. Another current interest is the design of casual creators, which are creativity support tools where the focus is on users having fun, rather than on efficient, professional production of artefacts. He is currently developing a casual creator for visual art called Art Done Quick for public release, which employs evolutionary and deep learning techniques to deliver a fun-first experience while users make decorative art pieces. Any project involving generative technologies is of interest to Simon. Research Areas: Game AI Game Audio and Music Game Design Computational Creativity Player Experience Casual Creators Generative Deep Learning s.colton@qmul.ac.uk Email Mastodon https://ccg.doc.gold.ac.uk/ccg_old/simoncolton/cv/ Other links Website LinkedIn BlueSky Github Themes Accessibility Creative Computing Game AI Game Audio Player Research - Previous Next

< Back Dr Carlos Gonzalez Diaz University of York iGGi Alum Carlos is finishing his PhD at the University of York. He holds an MSc in Serious Games at the University of Skövde (Sweden) and a BSc in Software Engineering (Spain). He is been closely connected with industry throughout his PhD, having worked in the last years for Microsoft Research, Sony Interactive Entertainment R&D, Musemio Ltd R&D and Goldsmiths, UoL; as well as done consulting for tech companies such as Unity Technologies. A description of Carlos' research: The purpose of my PhD research is to advance game technologies by democratising the use of ML techniques among non-experts through innovative tools and plugins for game engines. I developed ML specific visual scritping languages and used mixed-methods research approaches to understand how to better support developers in creating VR interactions and the challenges behind human-AI interaction. I had several technical jobs throughout my PhD, as my expertise is highly applicable in both industry and academia. Thanks to the broad range of expertise that I gathered through many years of industrial work and academic study, I can tackle the challenges emerging from the inter-disciplinary nature of modern work: where user psychology, immersive technology and artificial intelligence intersect. Please refer to my website for completely up-to-date information regarding publications. Feel free to reach out if you want more information or want to chat about my/your work. I am looking for positions starting on February 2023 onwards. carlos.gonzalezdiaz@york.ac.uk Email https://masto.ai/@carlotes247 Mastodon https://carlotes247.github.io Other links Website https://uk.linkedin.com/in/carlosglesdiaz LinkedIn BlueSky https://github.com/carlotes247 Github Supervisor(s): Prof. Sebastian Deterding Featured Publication(s): Embodied, in-medium design of VR game motion controls using interactive supervised learning Automatic Game Tuning for Strategic Diversity Programming by Moving: Interactive Machine Learning for Embodied Interaction Design InteractML: Node Based Tool to Empower Artists and Dancers in using Interactive Machine Learning for Designing Movement Interaction Movement interaction design for immersive media using interactive machine learning Using Machine Learning to Design Movement Interaction in Virtual Reality Interactive machine learning for more expressive game interactions Making Space for Social Time: Supporting Conversational Transitions Before, During, and After Video Meetings InteractML: Making machine learning accessible for creative practitioners working with movement interaction in immersive media Interactive Machine Learning for Embodied Interaction Design: A tool and methodology Bodystorming in SocialVR to Support Collaborative Embodied Ideation Themes Creative Computing Design & Development Game AI Immersive Technology https://www.youtube.com/watch?v=taVry9IQUjE https://www.youtube.com/watch?v=kkKU3MyBspM https://www.youtube.com/watch?v=WHiPav2l5gA Previous Next

< Back Remo Sasso Queen Mary University of London iGGi PG Researcher I hold a BSc and MSc in Artificial Intelligence at the University of Groningen (NL) and am currently a PhD student at the Queen Mary University of London under the supervision of Paulo Rauber. In addition to my academic work, I have worked as a Machine Learning engineer, and am currently the Head of AI at xDNA, an AI/Cybersecurity-based start-up from the Netherlands. Here I'm leading the initiative Project Aletheia, where we develop AI-driven tools to optimize the workflow of professional fact-checkers, with the overarching goal of ensuring information integrity in the world. Foundation World Models and Foundation Agents for Reinforcement Learning My research focuses on developing reinforcement learning algorithms that are both scalable and sample-efficient through Bayesian methods and model-based approaches, recently with a particular emphasis on Large Language Models (LLMs). My previous research focused on principled, efficient and scalable exploration algorithms for reinforcement learning, e.g. Poster Sampling for Deep Reinforcement Learning (ICML 2023), where we developed a reinforcement learning algorithm that can be considered state-of-the-art in Atari games. Currently I'm particularly interested in the integration of LLMs in the reinforcement learning framework, both as decision-making agents and simulators. My current research, called "Foundation World Models and Foundation Agents for Reinforcement Learning" investigates this integration in-depth and shows that large models show significant potential in various reinforcement learning tasks, ranging from decision-making in stochastic environments to serving as world models. r.sasso@qmul.ac.uk Email https://remosasso.github.io/ Mastodon Other links Website https://www.linkedin.com/in/remo-sasso-b9837a1ba/ LinkedIn BlueSky https://github.com/remosasso Github Supervisor: Dr Paulo Rauber Featured Publication(s): VDSC: Enhancing Exploration Timing with Value Discrepancy and State Counts Making Connections: Neurodevelopmental Changes in Brain Connectivity after Adverse Experiences in Early Adolescence Multi-Source Transfer Learning for Deep Model-Based Reinforcement Learning Simultaneous multi-view object recognition and grasping in open-ended domains Posterior Sampling for Deep Reinforcement Learning Themes Game AI - Previous Next

< Back Peyman Hosseini Queen Mary University of London iGGi PG Researcher Peyman is interested in using his computer science knowledge to support society's well-being. Raised in a family where almost everyone’s work is somehow related to mathematics and its applications, he became passionate about algorithms and combinatorics from an early age. This prompted him to pursue an undergraduate degree in computer engineering with a focus on IT and AI. This background led him to start his PhD at IGGI on building more powerful yet efficient Natural Language Processing models for analysing textual data, a rich and abundant source of gaming feedback. A description of Peyman's research: Peyman's research focuses on advancing deep learning architectures for natural language processing and building tools on top of state-of-the-art models. To contribute to the fundamental understanding and practical application of deep learning in natural language processing, focusing on efficiency and effectiveness, he pursues two main objectives: Designing more efficient models that match or surpass state-of-the-art performance with fewer parameters. Systematically analyzing language models to develop solutions that enhance their effectiveness for end-users, such as game studios. His recent accomplishments towards these goals include: 1. Developing novel attention mechanisms: 1.1 Optimized Attention: 25% parameter reduction 1.2 Efficient Attention: 50% parameter reduction 1.3 Super Attention: 25% parameter reduction with significant performance improvements in language and vision tasks 1.4 All mechanisms demonstrate comparable or superior performance to standard attention across various inputs. 2. Designing and training Hummingbird , a proof-of-concept small language model using Efficient Attention, available on HuggingFace. 3. Conducting a study on large language models' limitations in analyzing lengthy reviews for basic NLP tasks. Proposed solutions offer substantial performance improvements while reducing API costs by more than 90%. s.hosseini@qmul.ac.uk Email Mastodon https://peymanhosseini.net/ Other links Website https://www.linkedin.com/in/peyman-hosseini1 LinkedIn BlueSky https://github.com/Speymanhs Github Supervisors: Dr Ignacio Castro Prof. Matthew Purver Featured Publication(s): Cost-Effective Attention Mechanisms for Low Resource Settings: Necessity & Sufficiency of Linear Transformations Efficient solutions for an intriguing failure of llms: Long context window does not mean LLMs can analyze long sequences flawlessly Brain Drain Optimization (BRADO) Algorithm to Solve Multi-Objective Expert Team Formation Problem in Social Networks You Need to Pay Better Attention: Rethinking the Mathematics of Attention Mechanism GeoPos: A Minimal Positional Encoding for Enhanced Fine-Grained Details in Image Synthesis Using Convolutional Neural Networks Lon-ea at SemEval-2023 Task 11: A Comparison of Activation Functions for Soft and Hard Label Prediction GRACER: Improving the Accuracy of RACER Classifier Using A Greedy Approach Themes Game AI Player Research - Previous Next

< Back Alex Fletcher Queen Mary University of London iGGi Alum Alex Fletcher is a freelance audio engineer and junior game developer working on understanding the perceived flow and player experiences in mobile rhythm games and how a dynamic difficulty adjustment system would improve these experiences. The function of EEG and other biosensors as an additional measurement of player experience is of particular interest as further research in its use as an adaptive system. Other areas of research interest include game-based learning and games with a purpose. Please note: Updating of profile text in progress Email Mastodon Other links Website https://www.linkedin.com/in/alex-fletcher-64ab72176 LinkedIn BlueSky Github Themes Applied Games Game Audio Player Research - Previous Next

< Back Dr Marko Tot Queen Mary University of London iGGi Alum Hello! I'm Marko, and welcome to my page! As a part of the IGGI programme and Game AI research group, I'm working on adapting Statistical Forward Planning methods for complex environments. Statistical Forward Planning methods have proven to be effective in some simpler domains and, without requiring any prior learning, they provide a good out of the box AI algorithm. However, while these algorithms shine in certain games, they struggle to perform well in cases where the reward received from the game is sparse. In games where it takes a series of optimal actions to reach the goal, without any significant feedback from the environment in between, their performance drops significantly. My research is centered on solving this problem through automatic sub-goal generation and utilisation of local learned forward models. Creation of the sub-goals could be used to simulate the feedback from the environment and give regular rewards to the agent even in sparse and complex environments. I started my journey in video games when I got my first PC at the age of six, and at that point it was decided that I'm going to make a career out of it. So here I am, ~20 years later, a PhD. student at Queen Mary University of London, trying to make AI agents that can play games, and regularly spending too much time playing games under the excuse that it's all for 'research purpose'. m.tot@qmul.ac.uk Email Mastodon https://markotot.github.io/ Other links Website https://www.linkedin.com/in/markotot/ LinkedIn BlueSky https://github.com/markotot Github Supervisor(s): Dr Diego Pérez-Liébana Featured Publication(s): World and human action models towards gameplay ideation Turning Zeroes into Non-Zeroes: Sample Efficient Exploration with Monte Carlo Graph Search Making Something Out of Nothing: Monte Carlo Graph Search in Sparse Reward Environments What are you looking at? Team fight prediction through player camera Themes Game AI - Previous Next

< Back Dr Valerio Bonometti University of York iGGi Alum Game analytics and player psychology: creating reliable models of player motivation Motivation can be loosely defined as a process of the brain and the mind, capable of driving and deeply shaping human behaviour. Motivational processes are embedded in many everyday life situations, exerting their effects via a wide range of incentive mechanisms and objects. Understanding this process in a videogame context, however, requires a more holistic approach considering not just the incentive properties of the game but also the player personal characteristics. My project aims to create reliable cross-games models of player motivation taking into account the contribution of natural inter individual variability. This will be accomplished linking in-game behavioural data and psychological models via a hybrid approach, where findings from small scale experimental studies (hypothesis-driven) will guide the realization of large scale (data-driven) applications for predicting players' characteristics, future behaviour and motivation evolution. Being able to model player motivation and predict the trajectories of its evolution could possibly lead to personalized and dynamic engagement strategies able to adapt accordingly to the player characteristics and in-game behaviour. Achieving a similar goal would be of pivotal importance in industrial and gamification contexts. I obtained my bachelor degree in Psychological Science and my master degree in Clinical Psychology at Padova University (Italy). During my academic path I acquired knowledge in general psychology, cognitive psychology, psychophysiology, neuroscience and research methodology. After my master degree I spent a considerable amount of time as a research trainee, both abroad and in my country, always investigating the reward process and its effects in various contexts. During this period I worked on various projects across different fields ranging from psychophysiology, player research and game analytics. In my free time I enjoy practicing indoor climbing and travelling, I like figurative art in general and more specifically I’m a huge cinema and graphic-novel enthusiast. Supervisors: Prof. Anders Drachen, Dr Sam Devlin Please note: Updating of profile text in progress Email Mastodon Other links Website LinkedIn BlueSky Github Featured Publication(s): From Theory to Behaviour: Towards a General Model of Engagement Modelling early user-game interactions for joint estimation of survival time and churn probability Predicting skill learning in a large, longitudinal MOBA dataset Mind the gap: Distributed practice enhances performance in a MOBA game Approximating the Manifold Structure of Attributed Incentive Salience from Large-scale Behavioural Data: A Representation Learning Approach Based on Artificial Neural Networks Themes Player Research - Previous Next