Search Results

< Back iGGi Goldsmiths is located in New Cross, South East London, five miles from central London. iGGi is a consortium of four universities or sites: the University of York (York), Queen Mary University of London (QMUL), Goldsmiths, University of London (Goldsmiths), and the University of Essex (Essex). iGGi received funding in two phases: “iGGi 1” funded the first five cohorts of researchers across York, QMUL, Goldsmiths, and Essex and PGR intake spans from 2014 to 2018; “iGGi 2” started in 2019 with funding for a further five cohorts, this time only at York and QMUL. Due to either placement-related interruption of studies and/or switching to part-time studies, some of the "Goldsmiths iGGis" from the iGGi 1 funding round are still in the process of completing their PhD work. Goldsmiths is therefore still listed here as an active iGGi site, even though future iGGi main events such as the iGGi Conference and the iGGi Game Jam will take place at one of the iGGi 2 sites, i.e., either York or QMUL. iGGi Goldsmiths is part of Goldsmiths' Computing Department . You can find the Goldsmiths campus map in the gallery below. Goldsmiths, University of London (Goldsmiths) iGGi Goldmiths Gallery Main Building, Goldsmiths Campus Map, Goldsmiths Ben Pimlott Building, Goldsmiths Previous Next

Building Player Profiles in Mobile Monetisation: A Machine Learning Approach Theme Game Data Project proposed & supervised by David Zendle To discuss whether this project could become your PhD proposal please email: david.zendle@york.ac.uk < Back Building Player Profiles in Mobile Monetisation: A Machine Learning Approach Project proposal abstract: This project aims to use machine learning techniques to segment and profile mobile gamers in terms of their in-game spending. Estimates suggest that more than 2.6bn people play mobile games globally; that more than 80 billion mobile games are downloaded annually; and that mobile gaming accounts for almost $100bn in transactions every year. Despite the profitability of mobile gaming, little is known about how different kinds of players spend money in mobile games. Informal theories regarding specific differences in gaming are widely espoused: one influential model, for example, posits the existence of a small but profitable layer of heavily-involved 'whales', and much larger groups of smaller-spending 'dolphins' and 'minnows'. However, it is unclear whether this structure really does explain the monetisation of most games; and whether monetisation may vary between games; and between cultural contexts. In this project, we will take a data-driven approach, and apply a variety of machine learning techniques to large datasets of real player transactions. By both applying and developing algorithmic techniques for the analysis of such data, we will help build an understanding of how in-game spending may be profiled. This project would suit a machine learning specialist; a quantitative social scientist, or a data scientist wishing to do impactful work. It will be supervised by David Zendle, one of the world's leading experts on video game monetisation, and may involve one or more industrial partners who will share player data for the project. Supervisor: David Zendle Based at:

Evolving Perception for Game Agents Theme Game AI Project proposed & supervised by Alex Wade, Peter Cowling To discuss whether this project could become your PhD proposal please email: alex.wade@york.ac.uk < Back Evolving Perception for Game Agents Project proposal abstract: How does perception emerge? Hugely successful approaches to creating AI game playing agents such as MuZero, AlphaGo and AlphaStar learn the action to take in each state alongside a representation of the world to aid learning. For MuZero, AlphaGo and AlphaStar the representation is a prior distribution on how promising each move is in a given board position. The prior distribution can be seen as a highly effective way to perceive and simplify the game world, for greater decision-making fitness. In this project we will create game agents, for open world games such as Minecraft, which start from rudimentary sensors and simultaneously evolve a world representation while learning to make decisions leading to high fitness in the game world. We will investigate important scientific questions about how perception has evolved in humans, alongside creating interesting agents which might exhibit very weird and "alien" behaviours. Our internal representation of the world is conditioned both by evolution (for example, the physiology of the eye and brain) and also by learned experience. What sorts of perceptual systems might artificial agents develop in a simulated world? In this project we will develop simple 'open world' games into which we will release software agents with rudimentary sensory systems, possibly alongside human-controlled agents. These agents will be able to sense their world but not, initially, to perceive it (since perception is a combination of sensing and interpretation ). Both the sensory apparatus and the structure of the machine learning networks will be free to evolve (through genetic algorithms and reinforcement learning). Each generation will need to undergo a period of 'development' to train its networks on the current environment. We seek a motivated and talented student with a creative approach to research and skills in some of AI/machine learning, programming/game design, psychology/neuroscience and data analysis, and a willingness to learn new skills as necessary. Some travel to other international labs with an interest in this space may be possible. Supervisor: Alex Wade , Peter Cowling Based at:

< Back iGGi Con 2023 - Get Ready! Preparations for the next iGGi conference are underway! Better mark the date: iGGi Con 2023 13. + 14. September 2023 Queen Mary University of London This year's event will be packed with talks, workshops, panels, posters and more. For the first time ever, we will also run a Mini Expo with Industry stands. iGGi Con 2023 is a showcase for iGGi PGRs and friends, and iGGi Industry Partners as well as a networking platform for everyone interested/involved in the games industry and games research. So, don't miss out and REGISTER HERE TODAY !! Spaces are limited. Previous 3 Apr 2023 Next

< Back Mark the Date! iGGi Con 2024 - 11+12 Sep The iGGi Conference is the annual showcase of our 60+ PhD Researchers, allowing a birds-eye view into their work, and a chance to bring academic research, innovation and the games industry together. Following the success of the conferences in 2022 and 2023, the iGGi Con 2024 will take place at the University of York. More information to follow in a few months. Previous 20 Oct 2023 Next

< Back A community-sourced glossary of open scholarship terms Link Author(s) S Parsons, F Azevedo, MM Elsherif, S Guay, ON Shahim, GH Govaart, [...], N Ballou Abstract More info TBA Link

< Back Rolling horizon evolutionary algorithms for general video game playing Link Author(s) RD Gaina, S Devlin, SM Lucas, D Perez-Liebana Abstract More info TBA Link

< Back GAIG Meetup The recent Game AI Meetup took place on 01 March 2023. Talks and presentation included: Jakob Foerster (University of Oxford, UK): Opponent-Shaping and Interference in General-Sum Games Original talk abstract: In general-sum games, the interaction of self-interested learning agents commonly leads to collectively worst-case outcomes, such as defect-defect in the iterated prisoner's dilemma (IPD). To overcome this, some methods, such as Learning with Opponent-Learning Awareness (LOLA), shape their opponents' learning process. However, these methods are myopic since only a small number of steps can be anticipated, are asymmetric since they treat other agents as naive learners, and require the use of higher-order derivatives, which are calculated through white-box access to an opponent's differentiable learning algorithm. In this talk I will first introduce Model-Free Opponent Shaping (M-FOS), which overcomes all of these limitations. M-FOS learns in a meta-game in which each meta-step is an episode of the underlying (``inner'') game. The meta-state consists of the inner policies, and the meta-policy produces a new inner policy to be used in the next episode. M-FOS then uses generic model-free optimisation methods to learn meta-policies that accomplish long-horizon opponent shaping. I will finish off the talk with our recent results for adversarial (or cooperative) cheap-talk: How can agents interfere with (or support) the learning process of other agents without being able to act in the environment? Vanessa Volz ( modl.ai ): Establishing Trust in AI-based Tools for Game Development Original talk abstract: AI-based tools to support the game development process have long been a topic in Game AI research, with popular publications in testing, churn prediction, asset, level and even game generation. However, the adaptation of these techniques from the games industry has been hesitant at best: The small-scale and simplified examples researchers use to demonstrate their work understandably only seldom convince the industry to risk investing in AI tools. In this talk, I will speak about my experience establishing trust in AI-based tools to support creative processes in game development. Having worked on this topic in both industry and academia, I will address issues ranging from establishing a common language and explaining AI behaviour to issuing performance guarantees via benchmarking and theoretical analysis. Mike Preuss (Leiden University, The Netherlands): In the eye of the storm? Where are we going with game AI? Original talk abstract: Looking back at the last 10 years of research in Game AI we find that Big Tech research has shaken up things quite a lot. A number of challenges were resolved in record time (Go, StarCraft, etc) and AI algorithm development is probably still increasing in speed. However, it seems that the use of AI in game-making has not changed that much, and academic research often opts for "smaller problems", slowly turning towards Human-Centered AI as possibly most important general research direction. How can we approach the next leap predicted by Alex Champandard 10 years ago of really intelligent game AI? And where would we want that? Mike presents some inconclusive thoughts and ideas on future developments. The Game AI Meetup takes place several times a year. To sign up and receive updates, please register/join here: https://www.meetup.com/game-ai-meetup-gaim-of-london/ Previous 1 Mar 2023 Next

< Back iGGi Game Jam 2022 We thought that with summer fast approaching and the end of term in close sight, the time would be right to reflect back on some of iGGi’s more iGGi-ish events which took place earlier this year. One such event was the iGGi Game Jam . iGGi PGRs gather once a year to create a game from scratch in a limited space of time (usually over 48 hours). This is an opportunity for those less familiar with game design/development to experience the process first hand, for those who are already experienced and/or have worked in industry before to explore new tools and/or skills, but most of all, we look at it as a shared fun time dedicated to (re-)connecting within and across cohorts, socialising and exchanging ideas. Traditionally, the Game Jam is coincided with international online events such as the Global Game Jam or Ludum Dare. This year, however, all of the jamming iGGi groups opted out of submitting to the Global Game Jam (for which iGGi was a registered site) – partly out of protest over the Global Game Jam’s initial choice of sponsor, partly because many felt that a relaxed group atmosphere was preferable to the high-octane pressure that participation in a global competition brings with it. This is not to say that we didn’t succumb to competitive spirit: prizes in 5 different categories were given out iGGi-internally at the final presentations upon conclusion of the jam. The categories were Non-fungible Gameplay - Best mechanic and game experience Houston, We Have A Problem - Most successful fail in a making a game Best Buddy - Best multiplayer game I Just Can't Get Enough - Best storytelling, immersive or replayable experience Tech Neutral - Most original & climate friendly use of technology You can find the majority of the resulting mini-games/proofs of concept uploaded on Itch here: https://itch.io/jam/iggi22/entries Previous 30 Jan 2022 Next

< Back Predicting skill learning in a large, longitudinal MOBA dataset Link Author(s) M Aung, V Bonometti, A Drachen, P Cowling, AV Kokkinakis, C Yoder, ... Abstract More info TBA Link

< Back Unveiling modern board games: an ML-based approach to BoardGameGeek data analysis Link Author(s) D Nguyen, J Kritz, RD Gaina, D Perez-Liebana Abstract More info TBA Link

< Back iGGi at Develop:Brighton 2023 Following the previous year’s success, the iGGi stand featured again on the Expo floor of the 2023 Develop:Brighton conference . This year’s Expo saw a record breaking 4970 visitors over the two days of its running (43% up from 3471 in 2022), and many of these visitors dropped by the iGGi stand to chat to us. Shifts at the iGGi stand were covered by the conference-attending iGGi PG Researchers (15 in total) who appreciated a break from their (often solitary) thesis/research work. With its stand presence at the Develop:Brighton Expo, iGGi is promoting the intersection between games industry and research. We are looking to engage the two in conversation. We want to foster ways of knowledge exchange that are mutually beneficial to both sides. If you want to learn more about existing pathways of knowledge exchange between iGGi and games companies, you can check out our Industry Info + KE page , but we are also always open for new ideas and formats. Please do get in touch if you require any further information. Previous 14 Jul 2023 Next