Towards Generic Models of Player Experience

A touted use of Procedural Content Generation is generating content tailored to specific players. Previous work has relied on human identification of player profile features which are then mapped to level generator features. We present a machine-learned technique to train generators on Super Mario Bros. videos, generating levels based on latent play styles learned from the video. We evaluate the generators in comparison to the original levels and a machine-learned generator trained using simulated players.

Section: Related Workmentioning

confidence: 99%

Learning Player Tailored Content From Observation: Platformer Level Generation from Video Traces using LSTMs

Summerville

Guzdial

Mateas

et al. 2021

“…In general, these player modeling approaches have the drawback of being genre dependent. To address this, there has been some work on a general player modeling technique (Shaker, Shaker, and Abou-Zleikha 2015).…”

Section: Related Workmentioning

confidence: 99%

Personalized Procedural Content Generation for Increased Player Agency

2020

My interest is in the intersection between AI directors, player modeling, and procedural content generation. Using a combination of these tools, I believe that we can expand the space of tellable stories to create a story that is completely driven by the player’s actions in the game. There are two use cases for this work. The first is adding additional content to an open-world game by letting the player influence the world after the main story is completed, which would help address the problem of replayability. The second is to allow the player to create a completely personalized story from the beginning of the game, using the starting situation of the game as if it was the start of a simulation.

Modeling Player Engagement with Bayesian Hierarchical Models

Sawyer

Rowe

Azevedo

et al. 2018

Modeling player engagement is a key challenge in games. However, the gameplay signatures of engaged players can be highly context-sensitive, varying based on where the game is used or what population of players is using it. Traditionally, models of player engagement are investigated in a particular context, and it is unclear how effectively these models generalize to other settings and populations. In this work, we investigate a Bayesian hierarchical linear model for multi-task learning to devise a model of player engagement from a pair of datasets that were gathered in two complementary contexts: a Classroom Study with middle school students and a Laboratory Study with undergraduate students. Both groups of players used similar versions of Crystal Island, an educational interactive narrative game for science learning. Results indicate that the Bayesian hierarchical model outperforms both pooled and context-specific models in cross-validation measures of predicting player motivation from in-game behaviors, particularly for the smaller Classroom Study group. Further, we find that the posterior distributions of model parameters indicate that the coefficient for a measure of gameplay performance significantly differs between groups. Drawing upon their capacity to share information across groups, hierarchical Bayesian methods provide an effective approach for modeling player engagement with data from similar, but different, contexts.