CAIML #4
Our third event will focus on AI in games while gamescom - one of the world’s largest gaming events - is happening in Cologne. This time, we welcome you to STARTPLATZ to listen to four excellent speakers presenting their work on AI, machine learning, and data science for the gaming industry. As you can see below the applications of AI in games are manifold and we are looking forward to an exciting evening.
Join us to learn, network, discuss trends and innovations in the AI space, and exchange knowledge with AI experts, machine learning practitioners, and data scientists. A big thank you to our sponsors and event partners coparion and Gaffel for supporting this meetup.
Agenda
18:30 - Doors open & welcome
19:00 - Cristina Anselmi (Electronic Arts): Machine Translation at EA
In a world where everything needs to happen right away and people need to be able to communicate instantly, the gaming industry needs to adapt and make use of all the technology available to make this happen, not disregarding the importance of quality. Electronic Arts is investigating on how to best use computational linguistics and machine learning to cover these needs.
19:30 - Rafet Sifa (Fraunhofer IAIS): Matrix and Tensor Factorization for Behavioral Analysis in Games
Matrix and Tensor factorization models have been popular tools for learning representation in variety of descriptive, predictive and prescriptive machine learning tasks. Examples of their application areas include automatic feature learning, data clustering, link prediction and recommender systems. In this talk we will give an overview about applications of Matrix and Tensor factorization models for analyzing user behavior in computer games. We will cover the basic theory behind such representations, present methods to come up with proper models, discuss different ideas to enforce interpretability for human analysts and show real world use cases tackling variety of business intelligence problems.
20:00 - Break with drinks 🍻 and pizza 🍕 brought to you by coparion and Gaffel
20:30 - Ignacio Monereo (Google): Predicting your game’s monetization future
While there is no standard way of calculating LTV, all stakeholders in the mobile industry agree that this is a fundamental key metric to determine the success of a game or app. In this talk we cover some of the most popular ways of calculating LTV by top game developers and how these actors use this data in order to take decisions about their businesses.
21:00 - Christoffer Holmgård (modl.ai/Die Gute Fabrik/Northeastern University): AI Personas for Automatic Playtesting at Multiple Resolutions
Bots are useful for testing and evaluating game content and can even be used to control procedural content generation (PCG). These approaches are already seeing commercial use for both mobile and console games. Bots serve as stand-ins for human players and are used for level design support (e.g. for King’s Candy Crush) or for finding QA issues for humans to inspect further (e.g. through Sony’s Apollo system used with Horizon Zero Dawn). Not all players are the same, however: Players play in different styles and have different preferences. If we can capture and simulate these differences, we might get better automated testing, PCG, and synthetic analytics. In this talk, Christoffer Holmgård will talk about his work on simulating human playstyles for automated testing and procedural content generation using Procedural Personas.
21:30 - Closing & Networking