Playing strategy games is a challenging problem for artificial intelligence (AI). One of the major challenges is the large search space due to a diverse set of game components. In recent works, state abstraction has been applied to search-based game AI and has brought significant performance improvements. State abstraction techniques rely on reducing the search space, e.g., by aggregating similar states. However, the application of these abstractions is hindered because the quality of an abstraction is difficult to evaluate. Previous works hence abandon the abstraction in the middle of the search to not bias the search to a local optimum. This mechanism introduces a hyper-parameter to decide the time to abandon the current state abstraction. In this work, we propose a size-constrained state abstraction (SCSA), an approach that limits the maximum number of nodes being grouped together. We found that with SCSA, the abstraction is not required to be abandoned. Our empirical results on $3$ strategy games show that the SCSA agent outperforms the previous methods and yields robust performance over different games. Codes are open-sourced at \url{https://github.com/GAIGResearch/Stratega}.

Arguably, for the latter part of the late 20th and early 21st centuries, games have been seen as the drosophila of AI. Games are a set of exciting testbeds, whose solutions (in terms of identifying optimal players) would lead to machines that would possess some form of general intelligence, or at the very least help us gain insights toward building intelligent machines. Following impressive successes in traditional board games like Go, Chess, and Poker, but also video games like the Atari 2600 collection, it is clear that this is not the case. Games have been attacked successfully, but we are nowhere near AGI developments (or, as harsher critics might say, useful AI developments!). In this short vision paper, we argue that for game research to become again relevant to the AGI pathway, we need to be able to address \textit{Knightian uncertainty} in the context of games, i.e. agents need to be able to adapt to rapid changes in game rules on the fly with no warning, no previous data, and no model access.

Many enhancements to Monte-Carlo Tree Search (MCTS) have been proposed over almost two decades of general game playing and other artificial intelligence research. However, our ability to characterise and understand which variants work well or poorly in which games is still lacking. This paper describes work on an initial dataset that we have built to make progress towards such an understanding: 268,386 plays among 61 different agents across 1494 distinct games. We describe a preliminary analysis and work on training predictive models on this dataset, as well as lessons learned and future plans for a new and improved version of the dataset.

The news: A new AI agent from Google DeepMind can play different games, including ones it has never seen before such as Goat Simulator 3, a fun action game with exaggerated physics. Unlike earlier game-playing AI systems, which mastered only one game or could only follow single goals or commands, this new agent is able to play a variety of different games, including Valheim and No Man's Sky. How they did it: Researchers were able to get it to follow text commands to play seven different games and move around in three different 3D research environments. They trained it on lots of examples of humans playing video games, alongside keyboard and mouse input and annotations of what the players did. Then they used an AI technique called imitation learning to teach the agent to play games as humans would.

A Google DeepMind artificial intelligence model can play different open-world video games including No Man's Sky like a human, by watching video from a screen, which could be a step towards generally intelligent AIs that operate in the corporeal world. Playing video games has long been a way to test the progress of AI systems, such as Google DeepMind's AI mastery of chess or Go, but these games have obvious ways to win or lose, making it relatively straightforward to train an AI to succeed at them. Open-world games with extraneous information that can be ignored and more abstract objectives, such as Minecraft, are harder for AI systems to crack. Because the array of choices available in the games makes them a little more like normal life, they are thought to be an important stepping stone towards training AI agents that could do jobs in the real world, such as control robots, and artificial general intelligence. Now, researchers at Google DeepMind have developed an AI they call a Scalable Instructable Multiworld Agent, or SIMA, which can play nine different video games and virtual environments it hasn't seen before using just the video feed from the game.

In this work, we attempt to bridge the two fields of finite-agent and infinite-agent games, by studying how the optimal policies of agents evolve with the number of agents (population size) in mean-field games, an agent-centric perspective in contrast to the existing works focusing typically on the convergence of the empirical distribution of the population. To this end, the premise is to obtain the optimal policies of a set of finite-agent games with different population sizes. However, either deriving the closed-form solution for each game is theoretically intractable, training a distinct policy for each game is computationally intensive, or directly applying the policy trained in a game to other games is sub-optimal. We address these challenges through the Population-size-Aware Policy Optimization (PAPO). Our contributions are three-fold. First, to efficiently generate efficient policies for games with different population sizes, we propose PAPO, which unifies two natural options (augmentation and hypernetwork) and achieves significantly better performance. PAPO consists of three components: i) the population-size encoding which transforms the original value of population size to an equivalent encoding to avoid training collapse, ii) a hypernetwork to generate a distinct policy for each game conditioned on the population size, and iii) the population size as an additional input to the generated policy. Next, we construct a multi-task-based training procedure to efficiently train the neural networks of PAPO by sampling data from multiple games with different population sizes. Finally, extensive experiments on multiple environments show the significant superiority of PAPO over baselines, and the analysis of the evolution of the generated policies further deepens our understanding of the two fields of finite-agent and infinite-agent games.

There's a lot to love about digital gaming, like being able to shop for and download games instantly from my couch. I don't miss hunting for discs or the irritating whir as they spin in the drive. But sharing digital games is a mess. It's confusing and needlessly complicated. It's not unusual to find my entire household playing video games.

Behavior trees (BTs) are a popular method of modeling the behavior of NPCs and enemy AI and have found widespread use in a large number of commercial games. In this paper, rather than use BTs to model game-playing agents, we demonstrate their use for modeling game design agents, defining behaviors as executing content generation tasks rather than in-game actions. Similar to how traditional BTs enable modeling behaviors in a modular and dynamic manner, BTs for PCG enable simple subtrees for generating parts of levels to be combined modularly to form more complex trees for generating whole levels as well as generators that can dynamically vary the generated content. We demonstrate this approach by using BTs to model generators for Super Mario Bros., Mega Man and Metroid levels as well as dungeon layouts and discuss several ways in which this PCGBT paradigm could be applied and extended in the future.

Representing games through their pixels offers a promising approach for building general-purpose and versatile game models. While games are not merely images, neural network models trained on game pixels often capture differences of the visual style of the image rather than the content of the game. As a result, such models cannot generalize well even within similar games of the same genre. In this paper we build on recent advances in contrastive learning and showcase its benefits for representation learning in games. Learning to contrast images of games not only classifies games in a more efficient manner; it also yields models that separate games in a more meaningful fashion by ignoring the visual style and focusing, instead, on their content. Our results in a large dataset of sports video games containing 100k images across 175 games and 10 game genres suggest that contrastive learning is better suited for learning generalized game representations compared to conventional supervised learning. The findings of this study bring us closer to universal visual encoders for games that can be reused across previously unseen games without requiring retraining or fine-tuning.

AlphaZero has achieved impressive performance in deep reinforcement learning by utilizing an architecture that combines search and training of a neural network in self-play. Many researchers are looking for ways to reproduce and improve results for other games/tasks. However, the architecture is designed to learn from scratch, tabula rasa, accepting a cold-start problem in self-play. Recently, a warm-start enhancement method for Monte Carlo Tree Search was proposed to improve the self-play starting phase. It employs a fixed parameter $I^\prime$ to control the warm-start length. Improved performance was reported in small board games. In this paper we present results with an adaptive switch method. Experiments show that our approach works better than the fixed $I^\prime$, especially for "deep," tactical, games (Othello and Connect Four). We conjecture that the adaptive value for $I^\prime$ is also influenced by the size of the game, and that on average $I^\prime$ will increase with game size. We conclude that AlphaZero-like deep reinforcement learning benefits from adaptive rollout based warm-start, as Rapid Action Value Estimate did for rollout-based reinforcement learning 15 years ago.