Alexander Dockhorn

Linjie Xu, Jorge Hurtado-Grueso, Dominic Jeurissen et al.

Strategy video games challenge AI agents with their combinatorial search space caused by complex game elements. State abstraction is a popular technique that reduces the state space complexity. However, current state abstraction methods for games depend on domain knowledge, making their application to new games expensive. State abstraction methods that require no domain knowledge are studied extensively in the planning domain. However, no evidence shows they scale well with the complexity of strategy games. In this paper, we propose Elastic MCTS, an algorithm that uses state abstraction to play strategy games. In Elastic MCTS, the nodes of the tree are clustered dynamically, first grouped together progressively by state abstraction, and then separated when an iteration threshold is reached. The elastic changes benefit from efficient searching brought by state abstraction but avoid the negative influence of using state abstraction for the whole search. To evaluate our method, we make use of the general strategy games platform Stratega to generate scenarios of varying complexity. Results show that Elastic MCTS outperforms MCTS baselines with a large margin, while reducing the tree size by a factor of $10$. Code can be found at: https://github.com/egg-west/Stratega

Linjie Xu, Diego Perez-Liebana, Alexander Dockhorn

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 https://github.com/GAIGResearch/Stratega.

Aditya Mohan, Carolin Benjamins, Konrad Wienecke et al.

Although Reinforcement Learning (RL) has shown to be capable of producing impressive results, its use is limited by the impact of its hyperparameters on performance. This often makes it difficult to achieve good results in practice. Automated RL (AutoRL) addresses this difficulty, yet little is known about the dynamics of the hyperparameter landscapes that hyperparameter optimization (HPO) methods traverse in search of optimal configurations. In view of existing AutoRL approaches dynamically adjusting hyperparameter configurations, we propose an approach to build and analyze these hyperparameter landscapes not just for one point in time but at multiple points in time throughout training. Addressing an important open question on the legitimacy of such dynamic AutoRL approaches, we provide thorough empirical evidence that the hyperparameter landscapes strongly vary over time across representative algorithms from RL literature (DQN, PPO, and SAC) in different kinds of environments (Cartpole, Bipedal Walker, and Hopper) This supports the theory that hyperparameters should be dynamically adjusted during training and shows the potential for more insights on AutoRL problems that can be gained through landscape analyses. Our code can be found at https://github.com/automl/AutoRL-Landscape

Carlo Nübel, Alexander Dockhorn, Sanaz Mostaghim

Many works in the domain of artificial intelligence in games focus on board or video games due to the ease of reimplementing their mechanics. Decision-making problems in real-world sports share many similarities to such domains. Nevertheless, not many frameworks on sports games exist. In this paper, we present the tennis match simulation environment \textit{Match Point AI}, in which different agents can compete against real-world data-driven bot strategies. Next to presenting the framework, we highlight its capabilities by illustrating, how MCTS can be used in Match Point AI to optimize the shot direction selection problem in tennis. While the framework will be extended in the future, first experiments already reveal that generated shot-by-shot data of simulated tennis matches show realistic characteristics when compared to real-world data. At the same time, reasonable shot placement strategies emerge, which share similarities to the ones found in real-world tennis matches.

Ronja Fuchs, Florian Rupp, Timo Bertram et al.

Toxicity in online gaming communities remains a persistent challenge, manifesting across genres, platforms, and player interactions. While much research is focused on in-game toxicity, less is known about how toxic behavior varies between gaming communities on streaming platforms. To address this shortcoming, we analyze approximately 20 million chat messages from 4,452 streams, spanning seven game genres on Twitch. We categorize messages according to Twitch's toxicity taxonomy with a pre-trained Large Language Model using zero-shot classification. The taxonomy comprises four categories and eight subclasses, including harassment, discrimination, sexual content, and profanity. Our approach achieves an F1 score of 94.5% on the TextDetox dataset and demonstrates human-model agreement comparable to inter-human agreement. Our analysis reveals that 2.4% of all messages are classified as toxic, with notable differences across genres: streams of MOBA games exhibit the highest relative rate of toxicity (3.2%), and sports games show the lowest rate (2%). Furthermore, results indicate that individual games differ significantly in their toxicity distributions, even within genres, suggesting the existence of game-specific community norms and mechanics that shape toxic behavior beyond genre-level effects. These findings offer empirical insights into genre- and game-specific toxicity patterns on Twitch and can inform more targeted moderation strategies for gaming communities.

Linjie Xu, Zichuan Liu, Alexander Dockhorn et al.

One of the notorious issues for Reinforcement Learning (RL) is poor sample efficiency. Compared to single agent RL, the sample efficiency for Multi-Agent Reinforcement Learning (MARL) is more challenging because of its inherent partial observability, non-stationary training, and enormous strategy space. Although much effort has been devoted to developing new methods and enhancing sample efficiency, we look at the widely used episodic training mechanism. In each training step, tens of frames are collected, but only one gradient step is made. We argue that this episodic training could be a source of poor sample efficiency. To better exploit the data already collected, we propose to increase the frequency of the gradient updates per environment interaction (a.k.a. Replay Ratio or Update-To-Data ratio). To show its generality, we evaluate $3$ MARL methods on $6$ SMAC tasks. The empirical results validate that a higher replay ratio significantly improves the sample efficiency for MARL algorithms. The codes to reimplement the results presented in this paper are open-sourced at https://anonymous.4open.science/r/rr_for_MARL-0D83/.

Mohit Jiwatode, Alexander Dockhorn, Bodo Rosenhahn

Deep learning agents can achieve high performance in complex game domains without often understanding the underlying causal game mechanics. To address this, we investigate Causal Induction: the ability to infer governing laws from observational data, by tasking Large Language Models (LLMs) with reverse-engineering Video Game Description Language (VGDL) rules from gameplay traces. To reduce redundancy, we select nine representative games from the General Video Game AI (GVGAI) framework using semantic embeddings and clustering. We compare two approaches to VGDL generation: direct code generation from observations, and a two-stage method that first infers a structural causal model (SCM) and then translates it into VGDL. Both approaches are evaluated across multiple prompting strategies and controlled context regimes, varying the amount and form of information provided to the model, from just raw gameplay observations to partial VGDL specifications. Results show that the SCM-based approach more often produces VGDL descriptions closer to the ground truth than direct generation, achieving preference win rates of up to 81\% in blind evaluations and yielding fewer logically inconsistent rules. These learned SCMs can be used for downstream use cases such as causal reinforcement learning, interpretable agents, and procedurally generating novel but logically consistent games.

Raluca D. Gaina, Martin Balla, Alexander Dockhorn et al.

This document describes the design and implementation of the Tabletop Games framework (TAG), a Java-based benchmark for developing modern board games for AI research. TAG provides a common skeleton for implementing tabletop games based on a common API for AI agents, a set of components and classes to easily add new games and an import module for defining data in JSON format. At present, this platform includes the implementation of seven different tabletop games that can also be used as an example for further developments. Additionally, TAG also incorporates logging functionality that allows the user to perform a detailed analysis of the game, in terms of action space, branching factor, hidden information, and other measures of interest for Game AI research. The objective of this document is to serve as a central point where the framework can be described at length. TAG can be downloaded at: https://github.com/GAIGResearch/TabletopGames

Robin Schmöcker, Alexander Dockhorn, Bodo Rosenhahn

One approach to enhance Monte Carlo Tree Search (MCTS) is to improve its sample efficiency by grouping/abstracting states or state-action pairs and sharing statistics within a group. Though state-action pair abstractions are mostly easy to find in algorithms such as On the Go Abstractions in Upper Confidence bounds applied to Trees (OGA-UCT), nearly no state abstractions are found in either noisy or large action space settings due to constraining conditions. We provide theoretical and empirical evidence for this claim, and we slightly alleviate this state abstraction problem by proposing a weaker state abstraction condition that trades a minor loss in accuracy for finding many more abstractions. We name this technique Ideal Pruning Abstractions in UCT (IPA-UCT), which outperforms OGA-UCT (and any of its derivatives) across a large range of test domains and iteration budgets as experimentally validated. IPA-UCT uses a different abstraction framework from Abstraction of State-Action Pairs (ASAP) which is the one used by OGA-UCT, which we name IPA. Furthermore, we show that both IPA and ASAP are special cases of a more general framework that we call p-ASAP which itself is a special case of the ASASAP framework.

Ronja Fuchs, Robin Gieseke, Alexander Dockhorn

Balancing game difficulty in video games is a key task to create interesting gaming experiences for players. Mismatching the game difficulty and a player's skill or commitment results in frustration or boredom on the player's side, and hence reduces time spent playing the game. In this work, we explore balancing game difficulty using machine learning-based agents to challenge players based on their current behavior. This is achieved by a combination of two agents, in which one learns to imitate the player, while the second is trained to beat the first. In our demo, we investigate the proposed framework for personalized dynamic difficulty adjustment of AI agents in the context of the fighting game AI competition.

Mehmet Kayra Oğuz, Alexander Dockhorn

Markov Junior is a probabilistic programming language used for procedural content generation across various domains. However, its reliance on manually crafted and tuned probabilistic rule sets, also called grammars, presents a significant bottleneck, diverging from approaches that allow rule learning from examples. In this paper, we propose a novel solution to this challenge by introducing a genetic programming-based optimization framework for learning hierarchical rule sets automatically. Our proposed method ``Markov Senior'' focuses on extracting positional and distance relations from single input samples to construct probabilistic rules to be used by Markov Junior. Using a Kullback-Leibler divergence-based fitness measure, we search for grammars to generate content that is coherent with the given sample. To enhance scalability, we introduce a divide-and-conquer strategy that enables the efficient generation of large-scale content. We validate our approach through experiments in generating image-based content and Super Mario levels, demonstrating its flexibility and effectiveness. In this way, ``Markov Senior'' allows for the wider application of Markov Junior for tasks in which an example may be available, but the design of a generative rule set is infeasible.

Mohit Jiwatode, Leon Schlecht, Alexander Dockhorn

We propose RHEA CL, which combines Curriculum Learning (CL) with Rolling Horizon Evolutionary Algorithms (RHEA) to automatically produce effective curricula during the training of a reinforcement learning agent. RHEA CL optimizes a population of curricula, using an evolutionary algorithm, and selects the best-performing curriculum as the starting point for the next training epoch. Performance evaluations are conducted after every curriculum step in all environments. We evaluate the algorithm on the \textit{DoorKey} and \textit{DynamicObstacles} environments within the Minigrid framework. It demonstrates adaptability and consistent improvement, particularly in the early stages, while reaching a stable performance later that is capable of outperforming other curriculum learners. In comparison to other curriculum schedules, RHEA CL has been shown to yield performance improvements for the final Reinforcement learning (RL) agent at the cost of additional evaluation during training.

Robin Schmöcker, Lennart Kampmann, Alexander Dockhorn

One paradigm of Monte Carlo Tree Search (MCTS) improvements is to build and use state and/or action abstractions during the tree search. Non-exact abstractions, however, introduce an approximation error making convergence to the optimal action in the abstract space impossible. Hence, as proposed as a component of Elastic Monte Carlo Tree Search by Xu et al., abstraction algorithms should eventually drop the abstraction. In this paper, we propose two novel abstraction dropping schemes, namely OGA-IAAD and OGA-CAD which can yield clear performance improvements whilst being safe in the sense that the dropping never causes any notable performance degradations contrary to Xu's dropping method. OGA-IAAD is designed for time critical settings while OGA-CAD is designed to improve the MCTS performance with the same number of iterations.

Robin Schmöcker, Christoph Schnell, Alexander Dockhorn

The Upper Confidence Bounds For Trees (UCT) algorithm is not agnostic to the reward scale of the game it is applied to. For zero-sum games with the sparse rewards of $\{-1,0,1\}$ at the end of the game, this is not a problem, but many games often feature dense rewards with hand-picked reward scales, causing a node's Q-value to span different magnitudes across different games. In this paper, we evaluate various strategies for adaptively choosing the UCT exploration constant $λ$, called $λ$-strategies, that are agnostic to the game's reward scale. These $λ$-strategies include those proposed in the literature as well as five new strategies. Given our experimental results, we recommend using one of our newly suggested $λ$-strategies, which is to choose $λ$ as $2 \cdot σ$ where $σ$ is the empirical standard deviation of all state-action pairs' Q-values of the search tree. This method outperforms existing $λ$-strategies across a wide range of tasks both in terms of a single parameter value and the peak performances obtained by optimizing all available parameters.

Robin Schmöcker, Alexander Dockhorn, Bodo Rosenhahn

We introduce a novel, drop-in modification to Monte Carlo Tree Search's (MCTS) decision policy that we call AUPO. Comparisons based on a range of IPPC benchmark problems show that AUPO clearly outperforms MCTS. AUPO is an automatic action abstraction algorithm that solely relies on reward distribution statistics acquired during the MCTS. Thus, unlike other automatic abstraction algorithms, AUPO requires neither access to transition probabilities nor does AUPO require a directed acyclic search graph to build its abstraction, allowing AUPO to detect symmetric actions that state-of-the-art frameworks like ASAP struggle with when the resulting symmetric states are far apart in state space. Furthermore, as AUPO only affects the decision policy, it is not mutually exclusive with other abstraction techniques that only affect the tree search.

Robin Schmöcker, Alexander Dockhorn, Bodo Rosenhahn

One weakness of Monte Carlo Tree Search (MCTS) is its sample efficiency which can be addressed by building and using state and/or action abstractions in parallel to the tree search such that information can be shared among nodes of the same layer. The primary usage of abstractions for MCTS is to enhance the Upper Confidence Bound (UCB) value during the tree policy by aggregating visits and returns of an abstract node. However, this direct usage of abstractions does not take the case into account where multiple actions with the same parent might be in the same abstract node, as these would then all have the same UCB value, thus requiring a tiebreak rule. In state-of-the-art abstraction algorithms such as pruned On the Go Abstractions (pruned OGA), this case has not been noticed, and a random tiebreak rule was implicitly chosen. In this paper, we propose and empirically evaluate several alternative intra-abstraction policies, several of which outperform the random policy across a majority of environments and parameter settings.

Robin Schmöcker, Alexander Dockhorn, Bodo Rosenhahn

A core challenge of Monte Carlo Tree Search (MCTS) is its sample efficiency, which can be improved by grouping state-action pairs and using their aggregate statistics instead of single-node statistics. On the Go Abstractions in Upper Confidence bounds applied to Trees (OGA-UCT) is the state-of-the-art MCTS abstraction algorithm for deterministic environments that builds its abstraction using the Abstractions of State-Action Pairs (ASAP) framework, which aims to detect states and state-action pairs with the same value under optimal play by analysing the search graph. ASAP, however, requires two state-action pairs to have the same immediate reward, which is a rigid condition that limits the number of abstractions that can be found and thereby the sample efficiency. In this paper, we break with the paradigm of grouping value-equivalent states or state-action pairs and instead group states and state-action pairs with possibly different values as long as the difference between their values can be inferred. We call this abstraction framework Known Value Difference Abstractions (KVDA), which infers the value differences by analysis of the immediate rewards and modifies OGA-UCT to use this framework instead. The modification is called KVDA-UCT, which detects significantly more abstractions than OGA-UCT, introduces no additional parameter, and outperforms OGA-UCT on a variety of deterministic environments and parameter settings.

Manuel Eberhardinger, James Goodman, Alexander Dockhorn et al.

Large language models (LLMs) have shown impressive capabilities in generating program code, opening exciting opportunities for applying program synthesis to games. In this work, we explore the potential of LLMs to directly synthesize usable code for a wide range of gaming applications, focusing on two programming languages, Python and Java. We use an evolutionary hill-climbing algorithm, where the mutations and seeds of the initial programs are controlled by LLMs. For Python, the framework covers various game-related tasks, including five miniature versions of Atari games, ten levels of Baba is You, an environment inspired by Asteroids, and a maze generation task. For Java, the framework contains 12 games from the TAG tabletop games framework. Across 29 tasks, we evaluated 12 language models for Python and 8 for Java. Our findings suggest that the performance of LLMs depends more on the task than on model size. While larger models generate more executable programs, these do not always result in higher-quality solutions but are much more expensive. No model has a clear advantage, although on any specific task, one model may be better. Trying many models on a problem and using the best results across them is more reliable than using just one.

Alexander Dockhorn, Jorge Hurtado-Grueso, Dominik Jeurissen et al.

Portfolio methods represent a simple but efficient type of action abstraction which has shown to improve the performance of search-based agents in a range of strategy games. We first review existing portfolio techniques and propose a new algorithm for optimization and action-selection based on the Rolling Horizon Evolutionary Algorithm. Moreover, a series of variants are developed to solve problems in different aspects. We further analyze the performance of discussed agents in a general strategy game-playing task. For this purpose, we run experiments on three different game-modes of the Stratega framework. For the optimization of the agents' parameters and portfolio sets we study the use of the N-tuple Bandit Evolutionary Algorithm. The resulting portfolio sets suggest a high diversity in play-styles while being able to consistently beat the sample agents. An analysis of the agents' performance shows that the proposed algorithm generalizes well to all game-modes and is able to outperform other portfolio methods.

Diego Perez-Liebana, Cristina Guerrero-Romero, Alexander Dockhorn et al.

Designing agents that are able to achieve different play-styles while maintaining a competitive level of play is a difficult task, especially for games for which the research community has not found super-human performance yet, like strategy games. These require the AI to deal with large action spaces, long-term planning and partial observability, among other well-known factors that make decision-making a hard problem. On top of this, achieving distinct play-styles using a general algorithm without reducing playing strength is not trivial. In this paper, we propose Portfolio Monte Carlo Tree Search with Progressive Unpruning for playing a turn-based strategy game (Tribes) and show how it can be parameterized so a quality-diversity algorithm (MAP-Elites) is used to achieve different play-styles while keeping a competitive level of play. Our results show that this algorithm is capable of achieving these goals even for an extensive collection of game levels beyond those used for training.

Diego Perez-Liebana, Alexander Dockhorn, Jorge Hurtado Grueso et al.

Stratega, a general strategy games framework, has been designed to foster research on computational intelligence for strategy games. In contrast to other strategy game frameworks, Stratega allows to create a wide variety of turn-based and real-time strategy games using a common API for agent development. While the current version supports the development of turn-based strategy games and agents, we will add support for real-time strategy games in future updates. Flexibility is achieved by utilising YAML-files to configure tiles, units, actions, and levels. Therefore, the user can design and run a variety of games to test developed agents without specifically adjusting it to the game being generated. The framework has been built with a focus of statistical forward planning (SFP) agents. For this purpose, agents can access and modify game-states and use the forward model to simulate the outcome of their actions. While SFP agents have shown great flexibility in general game-playing, their performance is limited in case of complex state and action-spaces. Finally, we hope that the development of this framework and its respective agents helps to better understand the complex decision-making process in strategy games. Stratega can be downloaded at: https://github.research.its.qmul.ac.uk/eecsgameai/Stratega

Alexander Dockhorn, Simon M. Lucas, Vanessa Volz et al.

This paper examines learning approaches for forward models based on local cell transition functions. We provide a formal definition of local forward models for which we propose two basic learning approaches. Our analysis is based on the game Sokoban, where a wrong action can lead to an unsolvable game state. Therefore, an accurate prediction of an action's resulting state is necessary to avoid this scenario. In contrast to learning the complete state transition function, local forward models allow extracting multiple training examples from a single state transition. In this way, the Hash Set model, as well as the Decision Tree model, quickly learn to predict upcoming state transitions of both the training and the test set. Applying the model using a statistical forward planner showed that the best models can be used to satisfying degree even in cases in which the test levels have not yet been seen. Our evaluation includes an analysis of various local neighbourhood patterns and sizes to test the learners' capabilities in case too few or too many attributes are extracted, of which the latter has shown do degrade the performance of the model learner.

Alexander Dockhorn, Sanaz Mostaghim

The Hearthstone AI framework and competition motivates the development of artificial intelligence agents that can play collectible card games. A special feature of those games is the high variety of cards, which can be chosen by the players to create their own decks. In contrast to simpler card games, the value of many cards is determined by their possible synergies. The vast amount of possible decks, the randomness of the game, as well as the restricted information during the player's turn offer quite a hard challenge for the development of game-playing agents. This short paper introduces the competition framework and goes into more detail on the problems and challenges that need to be faced during the development process.

Simon M. Lucas, Alexander Dockhorn, Vanessa Volz et al.

This paper investigates the effect of learning a forward model on the performance of a statistical forward planning agent. We transform Conway's Game of Life simulation into a single-player game where the objective can be either to preserve as much life as possible or to extinguish all life as quickly as possible. In order to learn the forward model of the game, we formulate the problem in a novel way that learns the local cell transition function by creating a set of supervised training data and predicting the next state of each cell in the grid based on its current state and immediate neighbours. Using this method we are able to harvest sufficient data to learn perfect forward models by observing only a few complete state transitions, using either a look-up table, a decision tree or a neural network. In contrast, learning the complete state transition function is a much harder task and our initial efforts to do this using deep convolutional auto-encoders were less successful. We also investigate the effects of imperfect learned models on prediction errors and game-playing performance, and show that even models with significant errors can provide good performance.