Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2602.17213

Extending quantum theory with AI-assisted deterministic game theory

Florian Pauschitz, Ben Moseley, Ghislain Fourny

quant-phcs.AIcs.GTSubmitted Feb 19, 2026• Announced Feb 20, 2026• 0 citations
arXivHTMLPDF

Abstract

We present an AI-assisted framework for predicting individual runs of complex quantum experiments, including contextuality and causality (adaptive measurements), within our long-term programme of discovering a local hidden-variable theory that extends quantum theory. In order to circumvent impossibility theorems, we replace the assumption of free choice (measurement independence and parameter independence) with a weaker, compatibilistic version called contingent free choice. Our framework is based on interpreting complex quantum experiments as a Chess-like game between observers and the universe, which is seen as an economic agent minimizing action. The game structures corresponding to generic experiments such as fixed-causal-order process matrices or causal contextuality scenarios, together with a deterministic non-Nashian resolution algorithm that abandons unilateral deviation assumptions (free choice) and assumes Perfect Prediction instead, were described in previous work. In this new research, we learn the reward functions of the game, which contain a hidden variable, using neural networks. The cost function is the Kullback-Leibler divergence between the frequency histograms obtained through many deterministic runs of the game and the predictions of the extended Born rule. Using our framework on the specific case of the EPR 2-2-2 experiment acts as a proof-of-concept and a toy local-realist hidden-variable model that non-Nashian quantum theory is a promising avenue towards a local hidden-variable theory. Our framework constitutes a solid foundation, which can be further expanded in order to fully discover a complete quantum theory.

Comments
Extended abstract, 3 pages plus references. Preprint in progress

Related papers

Contingent Free Choice: On Extending Quantum Theory to a Contextual, Deterministic Theory With Improved Predictive Power

Ghislain Fourny

The non-extensibility of quantum theory into a theory with improved predictive power is based on a strong assumption of independent free choice, in which the physicists pick a measurement axis independently of anything that couldn't have been caused by their decision. Independent free choice is also at the core of the Nash equilibrium and classical game theory. A more recent line of game-theoretical research based on weakening free choice leads to non-trivial solution concepts with desirable properties such as at-most uniqueness, Pareto optimality, and contextuality. We show how introducing contingent free choice in the foundations of quantum theory yields a class of deterministic and contextual theories with an improved predictive power, and contrast them with the pilot-wave theory. Specifically, we suggest that quantum experiments, such as the EPR experiment, involving measurements located in spacetime, can be recast as dynamic games with imperfect information involving human agents and the universe. The underlying idea is that a physicist picking a measurement axis and the universe picking a measurement outcome are two faces of the same physical contingency phenomenon. The classical, Nashian resolution of these games based on independent free choice is analogous to local hidden variable theories, constrained by the Bell inequalities. On the other hand, in a setup in which agents are rational and omniscient in all possible worlds, under contingent free choice, the Perfectly Transparent Equilibrium provides a contextual resolution, based on the iterated elimination of inconsistent worlds, towards an at-most unique possible world, in which the outcomes of measurements that actually are carried out, and only them, are deterministically defined.

arXiv:1907.01962•Jul 3, 2019•3 citations

Nashian game theory is incompatible with quantum physics

Michał Bączyk, Ghislain Fourny

Abstract We suggest to look at quantum measurement outcomes not through the lens of probability theory, but instead through decision theory. We introduce an original game-theoretical framework, model and algorithmic procedure where measurement scenarios are multiplayer games with a structure all observers agree on. Measurement axes and, newly, measurement outcomes are modeled as decisions with nature being an action-minimizing economic agent. We translate physical notions of causality, correlation, counterfactuals, and contextuality to particular aspects of game theory. We investigate the causal consistency of dynamic games with imperfect information from the quantum perspective and conclude that counterfactual dependencies should be distinguished from causation and correlation as a separate phenomenon of its own. Most significantly, we observe that game theory based on Nash equilibria stands in contradiction with a violation of Bell inequalities. Hence, we propose that quantum physics should be analyzed with non-Nashian game theory, the inner workings of which we demonstrate using our proposed model.

arXiv:2112.03881•Quantum Studies Mathematics and Foundations•Sep 28, 2023•1 citation

Generating Massive-Scale Quantum Experiment Structures

Markus Bolinder

Quantum experiment structures provide a formal connection between causal models of quantum experiments and game-theoretic representations.Existing work in this area is predominantly theoretical, and explicit datasets and scalable software for studying these structures have been lacking.This thesis develops a computational framework for representing, generating, converting, and analyzing causal contextuality scenarios, spacetime games, and extensive form games.We present a Python package containing schema-based data representations and algorithms for generating consistent scenarios, constructing causally secured covers, converting causal contextuality scenarios to spacetime games, converting spacetime games to extensive form, and computing derived objects such as complete histories and reduced strategies.The implementation is integrated with Apache Spark for distributed execution, and with an automated testing pipeline that validates generated records and computes aggregate statistics.Using this framework, five annotated datasets are produced.The complete dataset for four measurements contains 6 233 520 structurally distinct scenarios after canonicalization, compared with 341 518 920 scenarios before eliminating equivalent causal structures.The sampled datasets include 4 194 304 scenarios with 5-15 measurements and 32 768 larger scenarios with 100-250 measurements.Where applicable, the datasets also contain corresponding spacetime games with histories and strategies, extensive form games, and deduplicated causal bridge representations.The results show that clean, stable, and causally secured covers occupy a small part of the full scenario space, but can be generated directly through constructive algorithms.They also show that computing reduced strategies and complete histories, together with recursive extensive form representations, imposes scalability limits.The resulting software and datasets provide a foundation for empirical investigations of quantum experiment structures at, previously unavailable, scales.

•Jan 1, 2026•0 citations

Views, variety and quantum mechanics

Lee Smolin

A non-local hidden variables theory for non-relativisitic quantum theory is presented, which gives a realist completion of quantum mechanics, in the sense of a complete description of individual events. The proposed fundamental theory is an extension of an energetic causal set theory, which assumes that time, events, causal structure, momentum and energy are fundamental. But space and the wave function are emergent. The beables of the theory are the views of the events, which are a subset of their causal pasts. Thus, this theory asserts that the universe is a causal network of events, which consists of partial views of itself as seen by looking backwards from each event. The fundamental dynamics is based on an action whose potential energy is proportional to the variety, which is a measure of the diversity of the views of the events, while the kinetic energy is proportional to its rate of change. The Schroedinger equation is derived to leading order in an expansion in density of the events of the fundamental histories. To higher order, there are computable corrections, non-linear in the wave function, from which new physical effects may be predicted.

arXiv:2105.03539•May 7, 2021•8 citations

Quantum game theory for 2 2 games: a mathematical framework

Gloria Ferraris, Veronica Umanità

We develop a rigorous mathematical framework for quantum game theory applied to static 2x2 games, extending classical concepts to the quantum setting where players may employ arbitrary unitary operations (pure strategies) or probability measures over the continuous group SU(2) (mixed strategies). The Eisert-Wilkens-Lewenstein protocol is introduced as the standard implementation of quantum 2x2 games. We prove the existence of Nash equilibria for continuous quantum mixed strategies via a fixed-point argument, generalising the classical Nash theorem to the quantum case.

quant-pharXiv:2605.15747•May 18, 2026•0 citations

Comment on Causal Networks and Freedom of Choice in Bell's Theorem

Marian Kupczyński

Bell inequalities may only be derived, if hidden variables do not depend on the experimental settings. The stochastic independence of hidden and setting variables is called: freedom of choice, free will, measurement independence or no conspiracy. By imbedding the Bell causal structure in a larger causal network the authors correctly prove, that one can explain and quantify possible violation of measurement independence without evoking the super-determinism. They assume the independence of the variables that causally determine the settings and investigate how they might become correlated with hidden variables. Using their extended causal networks they derive a contextual probabilistic model on which their further correct results are based. The authors seem to ignore that contextual probabilistic model may be derived directly using only probabilistic concepts and incorporating correctly setting dependent variables describing measuring instruments. In these contextual probabilistic models experimenters freedom of choice is not compromised and the results of Bell Tests including an apparent violation of Einsteinian nonsignaling may be explained in a locally causal way. Talking about freedom of choice is misleading and is rooted in incorrect understanding of Bayes Theorem. We explain why freedom of choice should be called noncontextuality and why its violation in Bell Tests confirms only the contextual character of quantum observables. Therefore, contextuality and not experimenters freedom of choice are important resources in quantum information.

arXiv:2201.08483•Jan 20, 2022•8 citations
We present an AI-assisted framework for predicting individual runs of complex quantum experiments, including contextuality and causality (adaptive measurements), within our long-term programme of discovering a local hidden-variable theory that extends quantum theory. In order to circumvent impossibility theorems, we replace the assumption of free choice (measurement independence and parameter independence) with a weaker, compatibilistic version called contingent free choice.
Our framework is based on interpreting complex quantum experiments as a Chess-like game between observers and the universe, which is seen as an economic agent minimizing action. The game structures corresponding to generic experiments such as fixed-causal-order process matrices or causal contextuality scenarios, together with a deterministic non-Nashian resolution algorithm that abandons unilateral deviation assumptions (free choice) and assumes Perfect Prediction instead, were described in previous work.
In this new research, we learn the reward functions of the game, which contain a hidden variable, using neural networks. The cost function is the Kullback-Leibler divergence between the frequency histograms obtained through many deterministic runs of the game and the predictions of the extended Born rule.
Using our framework on the specific case of the EPR 2-2-2 experiment acts as a proof-of-concept and a toy local-realist hidden-variable model that non-Nashian quantum theory is a promising avenue towards a local hidden-variable theory. Our framework constitutes a solid foundation, which can be further expanded in order to fully discover a complete quantum theory.
The non-extensibility of quantum theory into a theory with improved predictive power is based on a strong assumption of independent free choice, in which the physicists pick a measurement axis independently of anything that couldn't have been caused by their decision. Independent free choice is also at the core of the Nash equilibrium and classical game theory. A more recent line of game-theoretical research based on weakening free choice leads to non-trivial solution concepts with desirable properties such as at-most uniqueness, Pareto optimality, and contextuality. We show how introducing contingent free choice in the foundations of quantum theory yields a class of deterministic and contextual theories with an improved predictive power, and contrast them with the pilot-wave theory. Specifically, we suggest that quantum experiments, such as the EPR experiment, involving measurements located in spacetime, can be recast as dynamic games with imperfect information involving human agents and the universe. The underlying idea is that a physicist picking a measurement axis and the universe picking a measurement outcome are two faces of the same physical contingency phenomenon. The classical, Nashian resolution of these games based on independent free choice is analogous to local hidden variable theories, constrained by the Bell inequalities. On the other hand, in a setup in which agents are rational and omniscient in all possible worlds, under contingent free choice, the Perfectly Transparent Equilibrium provides a contextual resolution, based on the iterated elimination of inconsistent worlds, towards an at-most unique possible world, in which the outcomes of measurements that actually are carried out, and only them, are deterministically defined.
Abstract We suggest to look at quantum measurement outcomes not through the lens of probability theory, but instead through decision theory. We introduce an original game-theoretical framework, model and algorithmic procedure where measurement scenarios are multiplayer games with a structure all observers agree on. Measurement axes and, newly, measurement outcomes are modeled as decisions with nature being an action-minimizing economic agent. We translate physical notions of causality, correlation, counterfactuals, and contextuality to particular aspects of game theory. We investigate the causal consistency of dynamic games with imperfect information from the quantum perspective and conclude that counterfactual dependencies should be distinguished from causation and correlation as a separate phenomenon of its own. Most significantly, we observe that game theory based on Nash equilibria stands in contradiction with a violation of Bell inequalities. Hence, we propose that quantum physics should be analyzed with non-Nashian game theory, the inner workings of which we demonstrate using our proposed model.
Quantum experiment structures provide a formal connection between causal models of quantum experiments and game-theoretic representations.Existing work in this area is predominantly theoretical, and explicit datasets and scalable software for studying these structures have been lacking.This thesis develops a computational framework for representing, generating, converting, and analyzing causal contextuality scenarios, spacetime games, and extensive form games.We present a Python package containing schema-based data representations and algorithms for generating consistent scenarios, constructing causally secured covers, converting causal contextuality scenarios to spacetime games, converting spacetime games to extensive form, and computing derived objects such as complete histories and reduced strategies.The implementation is integrated with Apache Spark for distributed execution, and with an automated testing pipeline that validates generated records and computes aggregate statistics.Using this framework, five annotated datasets are produced.The complete dataset for four measurements contains 6 233 520 structurally distinct scenarios after canonicalization, compared with 341 518 920 scenarios before eliminating equivalent causal structures.The sampled datasets include 4 194 304 scenarios with 5-15 measurements and 32 768 larger scenarios with 100-250 measurements.Where applicable, the datasets also contain corresponding spacetime games with histories and strategies, extensive form games, and deduplicated causal bridge representations.The results show that clean, stable, and causally secured covers occupy a small part of the full scenario space, but can be generated directly through constructive algorithms.They also show that computing reduced strategies and complete histories, together with recursive extensive form representations, imposes scalability limits.The resulting software and datasets provide a foundation for empirical investigations of quantum experiment structures at, previously unavailable, scales.
A non-local hidden variables theory for non-relativisitic quantum theory is presented, which gives a realist completion of quantum mechanics, in the sense of a complete description of individual events. The proposed fundamental theory is an extension of an energetic causal set theory, which assumes that time, events, causal structure, momentum and energy are fundamental. But space and the wave function are emergent. The beables of the theory are the views of the events, which are a subset of their causal pasts. Thus, this theory asserts that the universe is a causal network of events, which consists of partial views of itself as seen by looking backwards from each event. The fundamental dynamics is based on an action whose potential energy is proportional to the variety, which is a measure of the diversity of the views of the events, while the kinetic energy is proportional to its rate of change. The Schroedinger equation is derived to leading order in an expansion in density of the events of the fundamental histories. To higher order, there are computable corrections, non-linear in the wave function, from which new physical effects may be predicted.
We develop a rigorous mathematical framework for quantum game theory applied to static 2x2 games, extending classical concepts to the quantum setting where players may employ arbitrary unitary operations (pure strategies) or probability measures over the continuous group SU(2) (mixed strategies). The Eisert-Wilkens-Lewenstein protocol is introduced as the standard implementation of quantum 2x2 games. We prove the existence of Nash equilibria for continuous quantum mixed strategies via a fixed-point argument, generalising the classical Nash theorem to the quantum case.
Bell inequalities may only be derived, if hidden variables do not depend on the experimental settings. The stochastic independence of hidden and setting variables is called: freedom of choice, free will, measurement independence or no conspiracy. By imbedding the Bell causal structure in a larger causal network the authors correctly prove, that one can explain and quantify possible violation of measurement independence without evoking the super-determinism. They assume the independence of the variables that causally determine the settings and investigate how they might become correlated with hidden variables. Using their extended causal networks they derive a contextual probabilistic model on which their further correct results are based. The authors seem to ignore that contextual probabilistic model may be derived directly using only probabilistic concepts and incorporating correctly setting dependent variables describing measuring instruments. In these contextual probabilistic models experimenters freedom of choice is not compromised and the results of Bell Tests including an apparent violation of Einsteinian nonsignaling may be explained in a locally causal way. Talking about freedom of choice is misleading and is rooted in incorrect understanding of Bayes Theorem. We explain why freedom of choice should be called noncontextuality and why its violation in Bell Tests confirms only the contextual character of quantum observables. Therefore, contextuality and not experimenters freedom of choice are important resources in quantum information.