arXiv subjects

Vojtěch Kovařík

Publications and source records attributed to Vojtěch Kovařík.

15 recordsLinked to original sources

arXiv2025

Multi-Agent Risks from Advanced AI

The rapid development of advanced AI agents and the imminent deployment of many instances of these agents will give rise to multi-agent systems of unprecedented complexity. These systems pose novel and under-explored risks. In this report, we provide a structured taxonomy of these risks by identifying three key failure modes (miscoordination, conflict, and collusion) based on agents' incentives, as well as seven key risk factors (information asymmetries, network effects, selection pressures, destabilising dynamics, commitment problems, emergent agency, and multi-agent security) that can underpin them. We highlight several important instances of each risk, as well as promising directions to help mitigate them. By anchoring our analysis in a range of real-world examples and experimental evidence, we illustrate the distinct challenges posed by multi-agent systems and their implications for the safety, governance, and ethics of advanced AI.

cs.MA↗

arXiv2025

Adapting Beyond the Depth Limit: Counter Strategies in Large Imperfect Information Games

We study the problem of adapting to a known sub-rational opponent during online play while remaining robust to rational opponents. We focus on large imperfect-information (zero-sum) games, which makes it impossible to inspect the whole game tree at once and necessitates the use of depth-limited search. However, all existing methods assume rational play beyond the depth-limit, which only allows them to adapt a very limited portion of the opponent's behaviour. We propose an algorithm Adapting Beyond Depth-limit (ABD) that uses a strategy-portfolio approach - which we refer to as matrix-valued states - for depth-limited search. This allows the algorithm to fully utilise all information about the opponent model, making it the first robust-adaptation method to be able to do so in large imperfect-information games. As an additional benefit, the use of matrix-valued states makes the algorithm simpler than traditional methods based on optimal value functions. Our experimental results in poker and battleship show that ABD yields more than a twofold increase in utility when facing opponents who make mistakes beyond the depth limit and also delivers significant improvements in utility and safety against randomly generated opponents.

cs.GT↗

arXiv2023

Revisiting Game Representations: The Hidden Costs of Efficiency in Sequential Decision-making Algorithms

Recent advancements in algorithms for sequential decision-making under imperfect information have shown remarkable success in large games such as limit- and no-limit poker. These algorithms traditionally formalize the games using the extensive-form game formalism, which, as we show, while theoretically sound, is memory-inefficient and computationally intensive in practice. To mitigate these challenges, a popular workaround involves using a specialized representation based on player specific information-state trees. However, as we show, this alternative significantly narrows the set of games that can be represented efficiently. In this study, we identify the set of large games on which modern algorithms have been benchmarked as being naturally represented by Sequential Bayesian Games. We elucidate the critical differences between extensive-form game and sequential Bayesian game representations, both theoretically and empirically. We further argue that the impressive experimental results often cited in the literature may be skewed, as they frequently stem from testing these algorithms only on this restricted class of games. By understanding these nuances, we aim to guide future research in developing more universally applicable and efficient algorithms for sequential decision-making under imperfect information.

cs.GT↗

arXiv2022

Value Functions for Depth-Limited Solving in Zero-Sum Imperfect-Information Games

We provide a formal definition of depth-limited games together with an accessible and rigorous explanation of the underlying concepts, both of which were previously missing in imperfect-information games. The definition works for an arbitrary extensive-form game and is not tied to any specific game-solving algorithm. Moreover, this framework unifies and significantly extends three approaches to depth-limited solving that previously existed in extensive-form games and multiagent reinforcement learning but were not known to be compatible. A key ingredient of these depth-limited games are value functions. Focusing on two-player zero-sum imperfect-information games, we show how to obtain optimal value functions and prove that public information provides both necessary and sufficient context for computing them. We provide a domain-independent encoding of the domains that allows for approximating value functions even by simple feed-forward neural networks, which are then able to generalize to unseen parts of the game. We use the resulting value network to implement a depth-limited version of counterfactual regret minimization. In three distinct domains, we show that the algorithm's exploitability is roughly linearly dependent on the value network's quality and that it is not difficult to train a value network with which depth-limited CFR's performance is as good as that of CFR with access to the full game.

cs.AI↗

arXiv2021

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↗

arXiv2021

(When) Is Truth-telling Favored in AI Debate?

For some problems, humans may not be able to accurately judge the goodness of AI-proposed solutions. Irving et al. (2018) propose that in such cases, we may use a debate between two AI systems to amplify the problem-solving capabilities of a human judge. We introduce a mathematical framework that can model debates of this type and propose that the quality of debate designs should be measured by the accuracy of the most persuasive answer. We describe a simple instance of the debate framework called feature debate and analyze the degree to which such debates track the truth. We argue that despite being very simple, feature debates nonetheless capture many aspects of practical debates such as the incentives to confuse the judge or stall to prevent losing. We then outline how these models should be generalized to analyze a wider range of debate phenomena.

cs.AI↗

arXiv2020

Solving Zero-Sum One-Sided Partially Observable Stochastic Games

Many security and other real-world situations are dynamic in nature and can be modelled as strictly competitive (or zero-sum) dynamic games. In these domains, agents perform actions to affect the environment and receive observations -- possibly imperfect -- about the situation and the effects of the opponent's actions. Moreover, there is no limitation on the total number of actions an agent can perform -- that is, there is no fixed horizon. These settings can be modelled as partially observable stochastic games (POSGs). However, solving general POSGs is computationally intractable, so we focus on a broad subclass of POSGs called one-sided POSGs. In these games, only one agent has imperfect information while their opponent has full knowledge of the current situation. We provide a full picture for solving one-sided POSGs: we (1) give a theoretical analysis of one-sided POSGs and their value functions, (2) show that a variant of a value-iteration algorithm converges in this setting, (3) adapt the heuristic search value-iteration algorithm for solving one-sided POSGs, (4) describe how to use approximate value functions to derive strategies in the game, and (5) demonstrate that our algorithm can solve one-sided POSGs of non-trivial sizes and analyze the scalability of our algorithm in three different domains: pursuit-evasion, patrolling, and search games.

cs.GT↗

arXiv2020

Absolute and non-absolute $\mathcal F$-Borel spaces

We investigate $\mathcal F$-Borel topological spaces. We focus on finding out how a~complexity of a~space depends on where the~space is embedded. Of a~particular interest is the~problem of determining whether a~complexity of given space $X$ is absolute (that is, the~same in every compactification of $X$). We show that the~complexity of metrizable spaces is absolute and provide a~sufficient condition for a~topological space to be absolutely $\mathcal F_{σδ}$. We then investigate the~relation between local and global complexity. To improve our understanding of $\mathcal F$-Borel spaces, we introduce different ways of representing an~$\mathcal F$-Borel set. We use these tools to construct a~hierarchy of $\mathcal F$-Borel spaces with non-absolute complexity, and to prove several other results.

math.GN↗

arXiv2019

Analysis of Hannan Consistent Selection for Monte Carlo Tree Search in Simultaneous Move Games

Hannan consistency, or no external regret, is a~key concept for learning in games. An action selection algorithm is Hannan consistent (HC) if its performance is eventually as good as selecting the~best fixed action in hindsight. If both players in a~zero-sum normal form game use a~Hannan consistent algorithm, their average behavior converges to a~Nash equilibrium (NE) of the~game. A similar result is known about extensive form games, but the~played strategies need to be Hannan consistent with respect to the~counterfactual values, which are often difficult to obtain. We study zero-sum extensive form games with simultaneous moves, but otherwise perfect information. These games generalize normal form games and they are a special case of extensive form games. We study whether applying HC algorithms in each decision point of these games directly to the~observed payoffs leads to convergence to a~Nash equilibrium. This learning process corresponds to a~class of Monte Carlo Tree Search algorithms, which are popular for playing simultaneous-move games but do not have any known performance guarantees. We show that using HC algorithms directly on the~observed payoffs is not sufficient to guarantee the~convergence. With an~additional averaging over joint actions, the~convergence is guaranteed, but empirically slower. We further define an~additional property of HC algorithms, which is sufficient to guarantee the~convergence without the~averaging and we empirically show that commonly used HC algorithms have this property.

cs.GT↗

arXiv2019

Problems with the EFG formalism: a solution attempt using observations

We argue that the extensive-form game (EFG) model isn't powerful enough to express all important aspects of imperfect information games, such as those related to decomposition and online game solving. We present a principled attempt to fix the formalism by considering information partitions that correspond to observations. We show that EFGs cannot be "fixed" without additional knowledge about the original task, and show how to extend the EFG model under this assumption. However, during our investigation, we ultimately concluded that a better solution is to abandon EFGs entirely and instead adopt a model based on partially observable stochastic games. The primary contribution of the text thus lies in exposing the problems with EFGs and presenting a detailed study on introducing imperfect information by adding observations into an underlying perfect-information model.

cs.GT↗

arXiv2018

Complexities and Representations of F-Borel Spaces

We investigate the $\mathcal F$-Borel complexity of topological spaces in their different compactifcations. We provide a simple proof of the fact that a space can have arbitrarily many different complexities in different compactifications. We also develop a theory of representations of $\mathcal F$-Borel sets, and show how to apply this theory to prove that the complexity of hereditarily Lindelöf spaces is absolute (that is, it is the same in every compactification). We use these representations to characterize the complexities attainable by a specific class of topological spaces. This provides an alternative proof of the first result, and implies the existence of a space with non-absolute additive complexity. We discuss the method used by Talagrand to construct the first example of a space with non-absolute complexity, hopefully providing an explanation which is more accessible than the original one. We also discuss the relation of complexity and local complexity, and show how to construct amalgamation-like compactifications.

math.GN↗

arXiv2018

Absolute $\mathcal F$-Borel classes

We investigate and compare $\mathcal F$-Borel classes and absolute $\mathcal F$-Borel classes. We provide precise examples distinguishing these two hierarchies. We also show that for separable metrizable spaces, $\mathcal F$-Borel classes are automatically absolute.

math.GN↗

arXiv2017

Absolute $F_{σδ}$ spaces

We prove that hereditarily Lindelöf space which is $F_{σδ}$ in some compactification is absolutely $F_{σδ}$. In particular, this implies that any separable Banach space is absolutely $F_{σδ}$ when equipped with the weak topology.

math.GN↗

arXiv2015

Analysis of Hannan Consistent Selection for Monte Carlo Tree Search in Simultaneous Move Games

Monte Carlo Tree Search (MCTS) has recently been successfully used to create strategies for playing imperfect-information games. Despite its popularity, there are no theoretic results that guarantee its convergence to a well-defined solution, such as Nash equilibrium, in these games. We partially fill this gap by analysing MCTS in the class of zero-sum extensive-form games with simultaneous moves but otherwise perfect information. The lack of information about the opponent's concurrent moves already causes that optimal strategies may require randomization. We present theoretic as well as empirical investigation of the speed and quality of convergence of these algorithms to the Nash equilibria. Primarily, we show that after minor technical modifications, MCTS based on any (approximately) Hannan consistent selection function always converges to an (approximate) subgame perfect Nash equilibrium. Without these modifications, Hannan consistency is not sufficient to ensure such convergence and the selection function must satisfy additional properties, which empirically hold for the most common Hannan consistent algorithms.

cs.GT↗

arXiv2013

Convergence of Monte Carlo Tree Search in Simultaneous Move Games

We study Monte Carlo tree search (MCTS) in zero-sum extensive-form games with perfect information and simultaneous moves. We present a general template of MCTS algorithms for these games, which can be instantiated by various selection methods. We formally prove that if a selection method is $ε$-Hannan consistent in a matrix game and satisfies additional requirements on exploration, then the MCTS algorithm eventually converges to an approximate Nash equilibrium (NE) of the extensive-form game. We empirically evaluate this claim using regret matching and Exp3 as the selection methods on randomly generated games and empirically selected worst case games. We confirm the formal result and show that additional MCTS variants also converge to approximate NE on the evaluated games.

cs.GT↗