Perttu Hämäläinen

Inan Evin, Perttu Hämäläinen, Christian Guckelsberger

Cutscenes form an integral part of many video games, but their creation is costly, time-consuming, and requires skills that many game developers lack. While AI has been leveraged to semi-automate cutscene production, the results typically lack the internal consistency and uniformity in style that is characteristic of professional human directors. We overcome this shortcoming with Cine-AI, an open-source procedural cinematography toolset capable of generating in-game cutscenes in the style of eminent human directors. Implemented in the popular game engine Unity, Cine-AI features a novel timeline and storyboard interface for design-time manipulation, combined with runtime cinematography automation. Via two user studies, each employing quantitative and qualitative measures, we demonstrate that Cine-AI generates cutscenes that people correctly associate with a target director, while providing above-average usability. Our director imitation dataset is publicly available, and can be extended by users and film enthusiasts.

Veera Vimpari, Annakaisa Kultima, Perttu Hämäläinen et al.

Text-to-image generation (TTIG) models, a recent addition to creative AI, can generate images based on a text description. These models have begun to rival the work of professional creatives, and sparked discussions on the future of creative work, loss of jobs, and copyright issues, amongst other important implications. To support the sustainable adoption of TTIG, we must provide rich, reliable and transparent insights into how professionals perceive, adopt and use TTIG. Crucially though, the public debate is shallow, narrow and lacking transparency, while academic work has focused on studying the use of TTIG in a general artist population, but not on the perceptions and attitudes of professionals in a specific industry. In this paper, we contribute a qualitative, exploratory interview study on TTIG in the Finnish videogame industry. Through a Template Analysis on semi-structured interviews with 14 game professionals, we reveal 12 overarching themes, structured into 49 sub-themes on professionals' perception, adoption and use of TTIG systems in games industry practice. Experiencing (yet another) change of roles and creative processes, our participants' reflections can inform discussions within the industry, be used by policymakers to inform urgently needed legislation, and support researchers in games, HCI and AI to support the sustainable, professional use of TTIG to benefit people and games as cultural artefacts.

Jeppe Theiss Kristensen, Christian Guckelsberger, Paolo Burelli et al.

The accurate and personalized estimation of task difficulty provides many opportunities for optimizing user experience. However, user diversity makes such difficulty estimation hard, in that empirical measurements from some user sample do not necessarily generalize to others. In this paper, we contribute a new approach for personalized difficulty estimation of game levels, borrowing methods from content recommendation. Using factorization machines (FM) on a large dataset from a commercial puzzle game, we are able to predict difficulty as the number of attempts a player requires to pass future game levels, based on observed attempt counts from earlier levels and levels played by others. In addition to performance and scalability, FMs offer the benefit that the learned latent variable model can be used to study the characteristics of both players and game levels that contribute to difficulty. We compare the approach to a simple non-personalized baseline and a personalized prediction using Random Forests. Our results suggest that FMs are a promising tool enabling game designers to both optimize player experience and learn more about their players and the game.

Nam Hee Kim, Jingjing May Liu, Jaakko Lehtinen et al.

We present the first motion generation system for playtesting virtual reality (VR) games. Our player model generates VR headset and handheld controller movements from in-game object arrangements, guided by style exemplars and aligned to maximize simulated gameplay score. We train on the large BOXRR-23 dataset and apply our framework on the popular VR game Beat Saber. The resulting model Robo-Saber produces skilled gameplay and captures diverse player behaviors, mirroring the skill levels and movement patterns specified by input style exemplars. Robo-Saber demonstrates promise in synthesizing rich gameplay data for predictive applications and enabling a physics-based whole-body VR playtesting agent.

Shaghayegh Roohi, Christian Guckelsberger, Asko Relas et al.

This paper presents a novel approach to automated playtesting for the prediction of human player behavior and experience. It has previously been demonstrated that Deep Reinforcement Learning (DRL) game-playing agents can predict both game difficulty and player engagement, operationalized as average pass and churn rates. We improve this approach by enhancing DRL with Monte Carlo Tree Search (MCTS). We also motivate an enhanced selection strategy for predictor features, based on the observation that an AI agent's best-case performance can yield stronger correlations with human data than the agent's average performance. Both additions consistently improve the prediction accuracy, and the DRL-enhanced MCTS outperforms both DRL and vanilla MCTS in the hardest levels. We conclude that player modelling via automated playtesting can benefit from combining DRL and MCTS. Moreover, it can be worthwhile to investigate a subset of repeated best AI agent runs, if AI gameplay does not yield good predictions on average.

Amin Babadi, Michiel van de Panne, C. Karen Liu et al.

We propose a novel method for exploring the dynamics of physically based animated characters, and learning a task-agnostic action space that makes movement optimization easier. Like several previous papers, we parameterize actions as target states, and learn a short-horizon goal-conditioned low-level control policy that drives the agent's state towards the targets. Our novel contribution is that with our exploration data, we are able to learn the low-level policy in a generic manner and without any reference movement data. Trained once for each agent or simulation environment, the policy improves the efficiency of optimizing both trajectories and high-level policies across multiple tasks and optimization algorithms. We also contribute novel visualizations that show how using target states as actions makes optimized trajectories more robust to disturbances; this manifests as wider optima that are easy to find. Due to its simplicity and generality, our proposed approach should provide a building block that can improve a large variety of movement optimization methods and applications.

Shaghayegh Roohi, Asko Relas, Jari Takatalo et al.

We propose a novel simulation model that is able to predict the per-level churn and pass rates of Angry Birds Dream Blast, a popular mobile free-to-play game. Our primary contribution is to combine AI gameplay using Deep Reinforcement Learning (DRL) with a simulation of how the player population evolves over the levels. The AI players predict level difficulty, which is used to drive a player population model with simulated skill, persistence, and boredom. This allows us to model, e.g., how less persistent and skilled players are more sensitive to high difficulty, and how such players churn early, which makes the player population and the relation between difficulty and churn evolve level by level. Our work demonstrates that player behavior predictions produced by DRL gameplay can be significantly improved by even a very simple population-level simulation of individual player differences, without requiring costly retraining of agents or collecting new DRL gameplay data for each simulated player.

Perttu Hämäläinen, Martin Trapp, Tuure Saloheimo et al.

We propose Deep Residual Mixture Models (DRMMs), a novel deep generative model architecture. Compared to other deep models, DRMMs allow more flexible conditional sampling: The model can be trained once with all variables, and then used for sampling with arbitrary combinations of conditioning variables, Gaussian priors, and (in)equality constraints. This provides new opportunities for interactive and exploratory machine learning, where one should minimize the user waiting for retraining a model. We demonstrate DRMMs in constrained multi-limb inverse kinematics and controllable generation of animations.

Perttu Hämäläinen, Juuso Toikka, Amin Babadi et al.

A large body of animation research focuses on optimization of movement control, either as action sequences or policy parameters. However, as closed-form expressions of the objective functions are often not available, our understanding of the optimization problems is limited. Building on recent work on analyzing neural network training, we contribute novel visualizations of high-dimensional control optimization landscapes; this yields insights into why control optimization is hard and why common practices like early termination and spline-based action parameterizations make optimization easier. For example, our experiments show how trajectory optimization can become increasingly ill-conditioned with longer trajectories, but parameterizing control as partial target states---e.g., target angles converted to torques using a PD-controller---can act as an efficient preconditioner. Both our visualizations and quantitative empirical data also indicate that neural network policy optimization scales better than trajectory optimization for long planning horizons. Our work advances the understanding of movement optimization and our visualizations should also provide value in educational use.

Amin Babadi, Kourosh Naderi, Perttu Hämäläinen

Animation and machine learning research have shown great advancements in the past decade, leading to robust and powerful methods for learning complex physically-based animations. However, learning can take hours or days, especially if no reference movement data is available. In this paper, we propose and evaluate a novel combination of techniques for accelerating the learning of stable locomotion movements through self-imitation learning of synthetic animations. First, we produce synthetic and cyclic reference movement using a recent online tree search approach that can discover stable walking gaits in a few minutes. This allows us to use reinforcement learning with Reference State Initialization (RSI) to find a neural network controller for imitating the synthesized reference motion. We further accelerate the learning using a novel curriculum learning approach called Termination Curriculum (TC), that adapts the episode termination threshold over time. The combination of the RSI and TC ensures that simulation budget is not wasted in regions of the state space not visited by the final policy. As a result, our agents can learn locomotion skills in just a few hours on a modest 4-core computer. We demonstrate this by producing locomotion movements for a variety of characters.

Perttu Hämäläinen, Amin Babadi, Xiaoxiao Ma et al.

Proximal Policy Optimization (PPO) is a highly popular model-free reinforcement learning (RL) approach. However, we observe that in a continuous action space, PPO can prematurely shrink the exploration variance, which leads to slow progress and may make the algorithm prone to getting stuck in local optima. Drawing inspiration from CMA-ES, a black-box evolutionary optimization method designed for robustness in similar situations, we propose PPO-CMA, a proximal policy optimization approach that adaptively expands the exploration variance to speed up progress. With only minor changes to PPO, our algorithm considerably improves performance in Roboschool continuous control benchmarks. Our results also show that PPO-CMA, as opposed to PPO, is significantly less sensitive to the choice of hyperparameters, allowing one to use it in complex movement optimization tasks without requiring tedious tuning.

Joose Rajamäki, Perttu Hämäläinen

We present a simple way to learn a transformation that maps samples of one distribution to the samples of another distribution. Our algorithm comprises an iteration of 1) drawing samples from some simple distribution and transforming them using a neural network, 2) determining pairwise correspondences between the transformed samples and training data (or a minibatch), and 3) optimizing the weights of the neural network being trained to minimize the distances between the corresponding vectors. This can be considered as a variant of the Iterative Closest Points (ICP) algorithm, common in geometric computer vision, although ICP typically operates on sensor point clouds and linear transforms instead of random sample sets and neural nonlinear transforms. We demonstrate the algorithm on simple synthetic data and MNIST data. We furthermore demonstrate that the algorithm is capable of handling distributions with both continuous and discrete variables.