Quantum common causes and quantum causal models

Abstract: Reichenbach's principle asserts that if two observed variables are found to be correlated, then there should be a causal explanation of these correlations. Furthermore, if the explanation is in terms of a common cause, then the conditional probability distribution over the variables given the complete common cause should factorize. The principle is generalized by the formalism of causal models, in which the causal relationships among variables constrain the form of their joint probability distribution. In the quantum case, however, the observed correlations in Bell experiments cannot be explained in the manner Reichenbach's principle would seem to demand. Motivated by this, we introduce a quantum counterpart to the principle. We demonstrate that under the assumption that quantum dynamics is fundamentally unitary, if a quantum channel with input A and outputs B and C is compatible with A being a complete common cause of B and C, then it must factorize in a particular way. Finally, we show how to generalize our quantum version of Reichenbach's principle to a formalism for quantum causal models, and provide examples of how the formalism works.

• The paper demonstrates that when a quantum channel factorizes as ρBC|A = ρB|A ρC|A, system A acts as a complete common cause.
• It employs unitary dilations and no-signaling principles to derive a quantum analogue of Reichenbach’s principle, establishing conditional independence.
• The findings impact quantum information theory and quantum machine learning by clarifying causal structures in quantum systems.

Quantum Common Causes and Quantum Causal Models

The paper "Quantum Common Causes and Quantum Causal Models" addresses a fundamental issue in quantum mechanics: the determination of causal relationships among quantum systems. The authors endeavor to generalize Reichenbach's principle, which in the classical context links statistical correlations among variables with underlying causal structures, to the quantum domain.

In classical statistics, Reichenbach's principle asserts that if two variables are correlated, there must be a causal explanation, which may be direct or through a common cause. In quantum mechanics, specifically in scenarios such as Bell experiments, observed correlations violate classical causal explanations. The quantum correlations in such experiments cannot be explained by a classical common cause model without invoking fine-tuning of parameters. This motivates the need for a quantum generalization of causal models.

The authors propose a quantum counterpart to Reichenbach's principle by focusing on channels where a quantum system $A$ acts as a common cause for systems $B$ and $C$. They introduce a condition for quantum conditional independence that serves as the quantum analogue of the classical condition $P(YZ|X) = P(Y|X)P(Z|X)$. Specifically, for a quantum channel $\rho_{BC|A}$, the outputs $B$ and $C$ are quantum conditionally independent given input $A$ if the channel factorizes as $\rho_{BC|A} = \rho_{B|A} \rho_{C|A}$.

The theoretical framework employed relies on the well-known quantum mechanics principle that any quantum evolution can be represented as unitary if one includes both the system and an environment. The authors utilize unitary dilations and leverage the property of no-signaling within quantum mechanics, positing that if a channel can be decomposed into independent influences, equivalent to classical conditional independence, then $A$ acts as a complete common cause for $B$ and $C$.

The paper's results are multi-faceted. Firstly, it is shown that if $\rho_{BC|A}$ satisfies the factorization condition $\rho_{BC|A} = \rho_{B|A} \rho_{C|A}$, then $A$ is a complete common cause. Secondly, there is an equivalent representation in terms of vanishing conditional mutual information $I(B:C|A) = 0$, highlighting entropic relationships as a parallel between classical and quantum causal models.

The implications of this work include a refined understanding of quantum causality, which has potential applications in quantum information theory and quantum machine learning. Quantum causal models could be instrumental in designing quantum algorithms for classical problems where quantum analogues prove advantageous. Additionally, these frameworks may lead to novel insights into quantum many-body systems and help uncover new quantum-classical separations.

The paper's theoretical approach draws on coherent interpretations of causal structure in quantum mechanics, opening avenues for further exploration. Future research could be directed towards extending these causal models to include multiple, interacting nodes and towards understanding quantum generalizations of causal inference principles. Insights from this work might contribute significantly to the overarching quest of reconciling quantum mechanics with relativistic causality.