arXiv subjects

Martin Schmid

Publications and source records attributed to Martin Schmid.

At least 19 recordsLinked to original sources

arXiv2026

Superhuman AI for Generals.io Using Self-Play Reinforcement Learning

We present a superhuman AI agent for Generals.io, a real-time strategy game that requires both long-horizon planning and short-term tactics under strong imperfect information. Trained for four days on 4x NVIDIA H200 GPUs, our agent reaches #1 on the public 1v1 leaderboard of over 5,000 human players, leading the second-ranked player by the same margin that separates second place from 25th, and beats the two top-ranked humans head-to-head with a combined 199-70 record across 269 ladder matches. A key enabler is a JAX-native simulator that reaches tens of millions of frames per second on a single GPU, roughly a 10,000x speedup over the prior simulator. On top of this, we train a vision transformer policy end-to-end by self-play with a policy-gradient loop and sparse win/loss reward, using top-advantage sample filtering and an exponential moving average of the policy parameters. Taken together, our findings highlight what matters, and what does not, once a fast simulator removes the data bottleneck.

cs.LG↗

arXiv2025

When and where higher-resolution climate data improve impact model performance

Climate impact assessments increasingly rely on high-resolution climate and forcing datasets, under the premise that finer detail enhances both the accuracy and policy relevance of projections. Yet systematic evaluations of when and where higher resolution actually improves impact model outcomes remain limited, and it is unclear whether increasing spatial resolution consistently enhances performance across sectors, regions, and forcing variables. Here we show that gains in climate input accuracy and impact model performance are largest when moving from coarse (60 km) to intermediate (10 km) resolution, while further refinement to 3 km and 1 km yields more modest and inconsistent benefits. Using cross-sectoral simulations from the Inter-Sectoral Impact Model Intercomparison Project, we find that higher resolution substantially improves model skill in temperature-sensitive impact models and topographically complex regions, whereas precipitation-driven and low-relief systems show weaker and less systematic improvements. For temperature, both climate inputs and model outputs improve most strongly at the 60 km to 10 km transition, with diminishing gains at finer scales; for precipitation, some models even exhibit reduced performance beyond 10 km. These results highlight that optimal resolution depends on sectoral and regional context, and point to the need for improving model process representation and downscaling techniques so that added spatial detail translates into meaningful skill gains. For data providers, this implies prioritizing resolutions that maximize improvements where they matter most, while for modelling groups and users it underscores the need for explicit benchmarking of resolution choices in climate impact assessments.

physics.ao-ph↗

arXiv2025

Massively Parallel Proof-Number Search for Impartial Games and Beyond

Proof-Number Search is a best-first search algorithm with many successful applications, especially in game solving. As large-scale computing clusters become increasingly accessible, parallelization is a natural way to accelerate computation. However, existing parallel versions of Proof-Number Search are known to scale poorly on many CPU cores. Using two parallelized levels and shared information among workers, we present the first massively parallel version of Proof-Number Search that scales efficiently even on a large number of CPUs. We apply our solver, enhanced with Grundy numbers for reducing game trees of impartial games, to the Sprouts game, a case study motivated by the long-standing Sprouts Conjecture. Our algorithm achieves 332.9$\times$ speedup on 1024 cores, significantly improving previous parallelizations and outperforming the state-of-the-art Sprouts solver GLOP by four orders of magnitude in runtime while generating proofs 1,000$\times$ more complex. Despite exponential growth in game tree size, our solver verified the Sprouts Conjecture for 42 new positions, nearly doubling the number of known outcomes.

cs.AI↗

arXiv2025

Neural Bayesian Filtering

We present Neural Bayesian Filtering (NBF), an algorithm for maintaining distributions over hidden states, called beliefs, in partially observable systems. NBF is trained to find a good latent representation of the beliefs induced by a task. It maps beliefs to fixed-length embedding vectors, which condition generative models for sampling. During filtering, particle-style updates compute posteriors in this embedding space using incoming observations and the environment's dynamics. NBF combines the computational efficiency of classical filters with the expressiveness of deep generative models - tracking rapidly shifting, multimodal beliefs while mitigating the risk of particle impoverishment. We validate NBF in state estimation tasks in three partially observable environments.

cs.LG↗

arXiv2025

SAM2RL: Towards Reinforcement Learning Memory Control in Segment Anything Model 2

Segment Anything Model 2 (SAM 2) has demonstrated strong performance in object segmentation tasks and has become the state-of-the-art for visual object tracking. The model stores information from previous frames in a memory bank, enabling temporal consistency across video sequences. Recent methods augment SAM 2 with hand-crafted update rules to better handle distractors, occlusions, and object motion. We propose a fundamentally different approach using reinforcement learning for optimizing memory updates in SAM 2 by framing memory control as a sequential decision-making problem. In an overfitting setup with a separate agent per video, our method achieves a relative improvement over SAM 2 that exceeds by more than three times the gains of existing heuristics. These results reveal the untapped potential of the memory bank and highlight reinforcement learning as a powerful alternative to hand-crafted update rules for memory control in visual object tracking.

cs.CV↗

arXiv2025

Artificial Generals Intelligence: Mastering Generals.io with Reinforcement Learning

We introduce a real-time strategy game environment based on Generals.io, a game with thousands of weekly active players. Our environment is fully compatible with Gymnasium and PettingZoo and is capable of running thousands of frames per second on commodity hardware. We also present a reference agent, trained with supervised pre-training and self-play, which reached the top 0.003% of the 1v1 human leaderboard after only 36 hours on a single H100 GPU. To accelerate learning, we incorporate potential-based reward shaping and memory features. Our contributions of a modular RTS benchmark and a competitive baseline agent provide an accessible yet challenging platform for advancing multi-agent reinforcement learning research. The documented code, together with examples and tutorials, is available at https://github.com/strakam/generals-bots.

cs.LG↗

arXiv2025

Meta-Learning in Self-Play Regret Minimization

Regret minimization is a general approach to online optimization which plays a crucial role in many algorithms for approximating Nash equilibria in two-player zero-sum games. The literature mainly focuses on solving individual games in isolation. However, in practice, players often encounter a distribution of similar but distinct games. For example, when trading correlated assets on the stock market, or when refining the strategy in subgames of a much larger game. Recently, offline meta-learning was used to accelerate one-sided equilibrium finding on such distributions. We build upon this, extending the framework to the more challenging self-play setting, which is the basis for most state-of-the-art equilibrium approximation algorithms for domains at scale. When selecting the strategy, our method uniquely integrates information across all decision states, promoting global communication as opposed to the traditional local regret decomposition. Empirical evaluation on normal-form games and river poker subgames shows our meta-learned algorithms considerably outperform other state-of-the-art regret minimization algorithms.

cs.GT↗

arXiv2024

Learning to Beat ByteRL: Exploitability of Collectible Card Game Agents

While Poker, as a family of games, has been studied extensively in the last decades, collectible card games have seen relatively little attention. Only recently have we seen an agent that can compete with professional human players in Hearthstone, one of the most popular collectible card games. Although artificial agents must be able to work with imperfect information in both of these genres, collectible card games pose another set of distinct challenges. Unlike in many poker variants, agents must deal with state space so vast that even enumerating all states consistent with the agent's beliefs is intractable, rendering the current search methods unusable and requiring the agents to opt for other techniques. In this paper, we investigate the strength of such techniques for this class of games. Namely, we present preliminary analysis results of ByteRL, the state-of-the-art agent in Legends of Code and Magic and Hearthstone. Although ByteRL beat a top-10 Hearthstone player from China, we show that its play in Legends of Code and Magic is highly exploitable.

cs.AI↗

arXiv2023

Learning not to Regret

The literature on game-theoretic equilibrium finding predominantly focuses on single games or their repeated play. Nevertheless, numerous real-world scenarios feature playing a game sampled from a distribution of similar, but not identical games, such as playing poker with different public cards or trading correlated assets on the stock market. As these similar games feature similar equilibra, we investigate a way to accelerate equilibrium finding on such a distribution. We present a novel "learning not to regret" framework, enabling us to meta-learn a regret minimizer tailored to a specific distribution. Our key contribution, Neural Predictive Regret Matching, is uniquely meta-learned to converge rapidly for the chosen distribution of games, while having regret minimization guarantees on any game. We validated our algorithms' faster convergence on a distribution of river poker games. Our experiments show that the meta-learned algorithms outpace their non-meta-learned counterparts, achieving more than tenfold improvements.

cs.GT↗

arXiv2021

Student of Games: A unified learning algorithm for both perfect and imperfect information games

Games have a long history as benchmarks for progress in artificial intelligence. Approaches using search and learning produced strong performance across many perfect information games, and approaches using game-theoretic reasoning and learning demonstrated strong performance for specific imperfect information poker variants. We introduce Student of Games, a general-purpose algorithm that unifies previous approaches, combining guided search, self-play learning, and game-theoretic reasoning. Student of Games achieves strong empirical performance in large perfect and imperfect information games -- an important step towards truly general algorithms for arbitrary environments. We prove that Student of Games is sound, converging to perfect play as available computation and approximation capacity increases. Student of Games reaches strong performance in chess and Go, beats the strongest openly available agent in heads-up no-limit Texas hold'em poker, and defeats the state-of-the-art agent in Scotland Yard, an imperfect information game that illustrates the value of guided search, learning, and game-theoretic reasoning.

cs.AI↗

arXiv2021

Search in Imperfect Information Games

From the very dawn of the field, search with value functions was a fundamental concept of computer games research. Turing's chess algorithm from 1950 was able to think two moves ahead, and Shannon's work on chess from $1950$ includes an extensive section on evaluation functions to be used within a search. Samuel's checkers program from 1959 already combines search and value functions that are learned through self-play and bootstrapping. TD-Gammon improves upon those ideas and uses neural networks to learn those complex value functions -- only to be again used within search. The combination of decision-time search and value functions has been present in the remarkable milestones where computers bested their human counterparts in long standing challenging games -- DeepBlue for Chess and AlphaGo for Go. Until recently, this powerful framework of search aided with (learned) value functions has been limited to perfect information games. As many interesting problems do not provide the agent perfect information of the environment, this was an unfortunate limitation. This thesis introduces the reader to sound search for imperfect information games.

cs.AI↗

arXiv2021

Sound Algorithms in Imperfect Information Games

Search has played a fundamental role in computer game research since the very beginning. And while online search has been commonly used in perfect information games such as Chess and Go, online search methods for imperfect information games have only been introduced relatively recently. This paper addresses the question of what is a sound online algorithm in an imperfect information setting of two-player zero-sum games. We argue that the~fixed-strategy~definitions of exploitability and $ε$-Nash equilibria are ill-suited to measure an online algorithm's worst-case performance. We thus formalize $ε$-soundness, a concept that connects the worst-case performance of an online algorithm to the performance of an $ε$-Nash equilibrium. As $ε$-soundness can be difficult to compute in general, we introduce a consistency framework -- a hierarchy that connects an online algorithm's behavior to a Nash equilibrium. These multiple levels of consistency describe in what sense an online algorithm plays "just like a fixed Nash equilibrium". These notions further illustrate the difference between perfect and imperfect information settings, as the same consistency guarantees have different worst-case online performance in perfect and imperfect information games. The definitions of soundness and the consistency hierarchy finally provide appropriate tools to analyze online algorithms in repeated imperfect information games. We thus inspect some of the previous online algorithms in a new light, bringing new insights into their worst-case performance guarantees.

cs.GT↗

arXiv2021

Solving Common-Payoff Games with Approximate Policy Iteration

For artificially intelligent learning systems to have widespread applicability in real-world settings, it is important that they be able to operate decentrally. Unfortunately, decentralized control is difficult -- computing even an epsilon-optimal joint policy is a NEXP complete problem. Nevertheless, a recently rediscovered insight -- that a team of agents can coordinate via common knowledge -- has given rise to algorithms capable of finding optimal joint policies in small common-payoff games. The Bayesian action decoder (BAD) leverages this insight and deep reinforcement learning to scale to games as large as two-player Hanabi. However, the approximations it uses to do so prevent it from discovering optimal joint policies even in games small enough to brute force optimal solutions. This work proposes CAPI, a novel algorithm which, like BAD, combines common knowledge with deep reinforcement learning. However, unlike BAD, CAPI prioritizes the propensity to discover optimal joint policies over scalability. While this choice precludes CAPI from scaling to games as large as Hanabi, empirical results demonstrate that, on the games to which CAPI does scale, it is capable of discovering optimal joint policies even when other modern multi-agent reinforcement learning algorithms are unable to do so. Code is available at https://github.com/ssokota/capi .

cs.AI↗

arXiv2020

The Advantage Regret-Matching Actor-Critic

Regret minimization has played a key role in online learning, equilibrium computation in games, and reinforcement learning (RL). In this paper, we describe a general model-free RL method for no-regret learning based on repeated reconsideration of past behavior. We propose a model-free RL algorithm, the AdvantageRegret-Matching Actor-Critic (ARMAC): rather than saving past state-action data, ARMAC saves a buffer of past policies, replaying through them to reconstruct hindsight assessments of past behavior. These retrospective value estimates are used to predict conditional advantages which, combined with regret matching, produces a new policy. In particular, ARMAC learns from sampled trajectories in a centralized training setting, without requiring the application of importance sampling commonly used in Monte Carlo counterfactual regret (CFR) minimization; hence, it does not suffer from excessive variance in large environments. In the single-agent setting, ARMAC shows an interesting form of exploration by keeping past policies intact. In the multiagent setting, ARMAC in self-play approaches Nash equilibria on some partially-observable zero-sum benchmarks. We provide exploitability estimates in the significantly larger game of betting-abstracted no-limit Texas Hold'em.

cs.AI↗

arXiv2020

Approximate exploitability: Learning a best response in large games

Researchers have demonstrated that neural networks are vulnerable to adversarial examples and subtle environment changes, both of which one can view as a form of distribution shift. To humans, the resulting errors can look like blunders, eroding trust in these agents. In prior games research, agent evaluation often focused on the in-practice game outcomes. While valuable, such evaluation typically fails to evaluate robustness to worst-case outcomes. Prior research in computer poker has examined how to assess such worst-case performance, both exactly and approximately. Unfortunately, exact computation is infeasible with larger domains, and existing approximations rely on poker-specific knowledge. We introduce ISMCTS-BR, a scalable search-based deep reinforcement learning algorithm for learning a best response to an agent, thereby approximating worst-case performance. We demonstrate the technique in several two-player zero-sum games against a variety of agents, including several AlphaZero-based agents.

cs.LG↗

arXiv2019

Low-Variance and Zero-Variance Baselines for Extensive-Form Games

Extensive-form games (EFGs) are a common model of multi-agent interactions with imperfect information. State-of-the-art algorithms for solving these games typically perform full walks of the game tree that can prove prohibitively slow in large games. Alternatively, sampling-based methods such as Monte Carlo Counterfactual Regret Minimization walk one or more trajectories through the tree, touching only a fraction of the nodes on each iteration, at the expense of requiring more iterations to converge due to the variance of sampled values. In this paper, we extend recent work that uses baseline estimates to reduce this variance. We introduce a framework of baseline-corrected values in EFGs that generalizes the previous work. Within our framework, we propose new baseline functions that result in significantly reduced variance compared to existing techniques. We show that one particular choice of such a function --- predictive baseline --- is provably optimal under certain sampling schemes. This allows for efficient computation of zero-variance value estimates even along sampled trajectories.

cs.GT↗

arXiv2019

Rethinking Formal Models of Partially Observable Multiagent Decision Making

Multiagent decision-making in partially observable environments is usually modelled as either an extensive-form game (EFG) in game theory or a partially observable stochastic game (POSG) in multiagent reinforcement learning (MARL). One issue with the current situation is that while most practical problems can be modelled in both formalisms, the relationship of the two models is unclear, which hinders the transfer of ideas between the two communities. A second issue is that while EFGs have recently seen significant algorithmic progress, their classical formalization is unsuitable for efficient presentation of the underlying ideas, such as those around decomposition. To solve the first issue, we introduce factored-observation stochastic games (FOSGs), a minor modification of the POSG formalism which distinguishes between private and public observation and thereby greatly simplifies decomposition. To remedy the second issue, we show that FOSGs and POSGs are naturally connected to EFGs: by "unrolling" a FOSG into its tree form, we obtain an EFG. Conversely, any perfect-recall timeable EFG corresponds to some underlying FOSG in this manner. Moreover, this relationship justifies several minor modifications to the classical EFG formalization that recently appeared as an implicit response to the model's issues with decomposition. Finally, we illustrate the transfer of ideas between EFGs and MARL by presenting three key EFG techniques -- counterfactual regret minimization, sequence form, and decomposition -- in the FOSG framework.

cs.AI↗

arXiv2019

Revisiting CFR+ and Alternating Updates

The CFR+ algorithm for solving imperfect information games is a variant of the popular CFR algorithm, with faster empirical performance on a range of problems. It was introduced with a theoretical upper bound on solution error, but subsequent work showed an error in one step of the proof. We provide updated proofs to recover the original bound.

cs.GT↗