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:2402.08060 · New Journal of Physics

Information gain and measurement disturbance for quantum agents

Arthur O. T. Pang, Noah Lupu-Gladstein, Y. Batuhan Yilmaz, C Pria Dobney, Rui Jie Tang, Aharon Brodutch, Aephraim M. Steinberg

Published Aug 25, 2026• 3 citations
arXivHTMLPDF

Abstract

Abstract The traditional formalism of quantum measurement (hereafter ‘TQM’) describes processes where some properties of quantum states are extracted and stored as classical information. While TQM is a natural and appropriate description of how humans interact with quantum systems, it is silent on the question of how a more general, quantum , agent would do so. How do we describe the observation of a system by an agent with the ability to store not only classical information but quantum states in its memory? In this paper, we extend the idea of measurement to a more general class of sensors for quantum agents which interact with a system in such a way that the agent’s memory stores information (classical or quantum) about the system under study. For appropriate sensory interactions, the quantum agent may ‘learn’ more about the system than would be possible under any set of classical measurements—but as we show, this comes at the cost of additional measurement disturbance. We experimentally demonstrate such a system and characterize the tradeoffs by considering the channel capacity required to erase the effect of a measurement.

DOI
10.1088/1367-2630/ae9e64

Related papers

Do qubits dream of entangled sheep? Quantum measurement without classical output

Noah Lupu-Gladstein, Aharon Brodutch, Hugo Ferretti +4 more

Abstract Quantum mechanics is usually formulated with an implicit assumption that agents who can observe and interact with the world are external to it and have a classical memory. However, there is no accepted way to define the quantum–classical cut and no a priori reason to rule out fully quantum agents with coherent quantum memories. In this work, we introduce an entirely quantum notion of measurement, called a sensation, to account for quantum agents that experience the world through quantum sensors. Sensations eschew probabilities and instead describe a deterministic flow of quantum information. We quantify the information gain and disturbance of a sensation using concepts from quantum information theory and find that sensations always disturb at least as much as they inform. Viewing measurements as sensations could lead to a new understanding of quantum theory in general and to new results in the context of quantum networks.

arXiv:2008.10617•New Journal of Physics•May 1, 2024•6 citations

Quantum Sensors and Actuators for Quantum Agents

Arthur O. T. Pang

Traditional quantum measurements describe a process in which properties of quantum states are converted and stored into some classical memory such that entities that interact classically, such as humans, can examine and interpret such information. However, these measurements are not inclusive of observers that possess quantum memory and the ``observations'' they may wish to make. Such observers (hereafter ``quantum agents'') may use sensors that, although outside the purview of traditional quantum measurement, nonetheless generate correlations between their memory and the state of interest. In this thesis, I presented work studying sensors for quantum agents as an extension of sensors given by the traditional formulation of quantum measurements. For example, the use of SWAP-like unitaries as sensors will put the state of interest directly into the quantum memory of our agent, but at the same time is not described by traditional quantum measurement. I study these quantum sensors, which can be more powerful sensors than those given by traditional quantum measurement, but at the same time give a stronger measurement disturbance than sensors described in traditional quantum measurements. Additionally, I present work studying a quantum actuator for quantum agents controlling the use of quantum channels. I studied how putting two channels in superpositions can result in noise channels being able to transmit more information than when the channels are used individually. I show that putting channels into superposition will allow access to an additional degree of freedom in the channel, which causes an increase in the information transmitted. Finally, I present a work unrelated to quantum agents exploring an ontological theory derived from Bohmian mechanics' assumptions but with momentum as the ontological variable. In Bohmian theory, the choice of position as the ontological variable is neither self-evident nor obvious, and Bohm-like theories with a different choice of the ontological variable can also be constructed. I present an experimental comparison of Bohmian theory and a momentum-ontologic Bohmian-like theory and argue for the importance and need for making and justifying the choice of ontology.

•Jun 1, 2024•0 citations

Quantum learning with a single-atom sensor

Yin Mo, Emilio Bagan, Giulio Chiribella

The ability to gather information and to act upon it is at the core of every learning agent. But what is the impact of quantum mechanics on an agent's ability to sense external inputs and to translate them into actions? Here we address the question for a prototype task of learning agency at the quantum scale: rotating a single spin based on information gathered by a single atom. We determine the ultimate performance limit for this task, revealing a fundamental tradeoff between entanglement at the sensing stage and coherence at the action stage: if the single-atom sensor is not entangled with the quantum system serving as the agent's internal memory, then the best learning strategy requires a coherent transfer of quantum information from the sensor to the system that controls the agent's actions. In contrast, if the sensor is initially entangled with the agent's memory, then the transfer of quantum information is no longer necessary. Our results indicate that the quantum properties of the sensor radically affect the optimal way to convert external stimuli into actions, revealing a link between quantum sensing and the behavior of quantum agents.

quant-pharXiv:2606.15071•Jun 16, 2026•0 citations

Relating measurement disturbance, information, and orthogonality

Yizhou Liu, John B. DeBrota

In the general theory of quantum measurement, one associates a positive-semidefinite operator on a $d$-dimensional Hilbert space to each of the $n$ possible outcomes of an arbitrary measurement. In the special case of a projective measurement, these operators are pairwise Hilbert-Schmidt orthogonal, but when $n>d$, orthogonality is restricted by positivity. This restriction allows us to more precisely state the quantum adage: information gain of a system is always accompanied by unavoidable disturbance. Specifically, we investigate three properties of a measurement with L\"uders rule updating: its disturbance, a measure of how the expected postmeasurement state deviates from the input; its measurement strength, a measure of the intrinsic information producing capacity of the measurement; and its orthogonality, a measure of the degree to which the measurement operators differ from an orthonormal set. These quantities satisfy an information-disturbance tradeoff relation that highlights the additional role played by orthogonality. Finally, we assess several classes of measurements on these grounds and identify symmetric informationally complete quantum measurements as the unique quantum analogs of a perfectly informative and nondisturbing classical ideal measurement.

arXiv:2105.02074•Physical Review A•Nov 22, 2021•2 citations

Quantum channels and memory effects

Filippo Caruso, Vittorio Giovannetti, Cosmo Lupo +1 more

Any physical process can be represented as a quantum channel mapping an initial state to a final state. Hence it can be characterized from the point of view of communication theory, i.e., in terms of its ability to transfer information. Quantum information provides a theoretical framework and the proper mathematical tools to accomplish this. In this context the notion of codes and communication capacities have been introduced by generalizing them from the classical Shannon theory of information transmission and error correction. The underlying assumption of this approach is to consider the channel not as acting on a single system, but on sequences of systems, which, when properly initialized allow one to overcome the noisy effects induced by the physical process under consideration. While most of the work produced so far has been focused on the case in which a given channel transformation acts identically and independently on the various elements of the sequence (memoryless configuration in jargon), correlated error models appear to be a more realistic way to approach the problem. A slightly different, yet conceptually related, notion of correlated errors applies to a single quantum system which evolves continuously in time under the influence of an external disturbance which acts on it in a non-Markovian fashion. This leads to the study of memory effects in quantum channels: a fertile ground where interesting novel phenomena emerge at the intersection of quantum information theory and other branches of physics. A survey is taken of the field of quantum channels theory while also embracing these specific and complex settings.

arXiv:1207.5435•Reviews of Modern Physics•Dec 10, 2014•349 citations

Reinforcement learning for quantum processes with memory

Josep Lumbreras, Ruo Cheng Huang, Yanglin Hu +2 more

In reinforcement learning, an agent interacts sequentially with an environment to maximize a reward, receiving only partial, probabilistic feedback. This creates a fundamental exploration-exploitation trade-off: the agent must explore to learn the hidden dynamics while exploiting this knowledge to maximize its target objective. While extensively studied classically, applying this framework to quantum systems requires dealing with hidden quantum states that evolve via unknown dynamics. We formalize this problem via a framework where the environment maintains a hidden quantum memory evolving via unknown quantum channels, and the agent intervenes sequentially using quantum instruments. For this setting, we adapt an optimistic maximum-likelihood estimation algorithm. We extend the analysis to continuous action spaces, allowing us to model general positive operator-valued measures (POVMs). By controlling the propagation of estimation errors through quantum channels and instruments, we prove that the cumulative regret of our strategy scales as $\widetilde{\mathcal{O}}(\sqrt{K})$ over $K$ episodes. Furthermore, via a reduction to the multi-armed quantum bandit problem, we establish information-theoretic lower bounds demonstrating that this sublinear scaling is strictly optimal up to polylogarithmic factors. As a physical application, we consider state-agnostic work extraction. When extracting free energy from a sequence of non-i.i.d. quantum states correlated by a hidden memory, any lack of knowledge about the source leads to thermodynamic dissipation. In our setting, the mathematical regret exactly quantifies this cumulative dissipation. Using our adaptive algorithm, the agent uses past energy outcomes to improve its extraction protocol on the fly, achieving sublinear cumulative dissipation, and, consequently, an asymptotically zero dissipation rate.

quant-pharXiv:2603.25138•Mar 27, 2026•3 citations
Abstract The traditional formalism of quantum measurement (hereafter ‘TQM’) describes processes where some properties of quantum states are extracted and stored as classical information. While TQM is a natural and appropriate description of how humans interact with quantum systems, it is silent on the question of how a more general, quantum , agent would do so. How do we describe the observation of a system by an agent with the ability to store not only classical information but quantum states in its memory? In this paper, we extend the idea of measurement to a more general class of sensors for quantum agents which interact with a system in such a way that the agent’s memory stores information (classical or quantum) about the system under study. For appropriate sensory interactions, the quantum agent may ‘learn’ more about the system than would be possible under any set of classical measurements—but as we show, this comes at the cost of additional measurement disturbance. We experimentally demonstrate such a system and characterize the tradeoffs by considering the channel capacity required to erase the effect of a measurement.
Abstract Quantum mechanics is usually formulated with an implicit assumption that agents who can observe and interact with the world are external to it and have a classical memory. However, there is no accepted way to define the quantum–classical cut and no a priori reason to rule out fully quantum agents with coherent quantum memories. In this work, we introduce an entirely quantum notion of measurement, called a sensation, to account for quantum agents that experience the world through quantum sensors. Sensations eschew probabilities and instead describe a deterministic flow of quantum information. We quantify the information gain and disturbance of a sensation using concepts from quantum information theory and find that sensations always disturb at least as much as they inform. Viewing measurements as sensations could lead to a new understanding of quantum theory in general and to new results in the context of quantum networks.
Traditional quantum measurements describe a process in which properties of quantum states are converted and stored into some classical memory such that entities that interact classically, such as humans, can examine and interpret such information. However, these measurements are not inclusive of observers that possess quantum memory and the ``observations'' they may wish to make. Such observers (hereafter ``quantum agents'') may use sensors that, although outside the purview of traditional quantum measurement, nonetheless generate correlations between their memory and the state of interest. In this thesis, I presented work studying sensors for quantum agents as an extension of sensors given by the traditional formulation of quantum measurements. For example, the use of SWAP-like unitaries as sensors will put the state of interest directly into the quantum memory of our agent, but at the same time is not described by traditional quantum measurement. I study these quantum sensors, which can be more powerful sensors than those given by traditional quantum measurement, but at the same time give a stronger measurement disturbance than sensors described in traditional quantum measurements. Additionally, I present work studying a quantum actuator for quantum agents controlling the use of quantum channels. I studied how putting two channels in superpositions can result in noise channels being able to transmit more information than when the channels are used individually. I show that putting channels into superposition will allow access to an additional degree of freedom in the channel, which causes an increase in the information transmitted. Finally, I present a work unrelated to quantum agents exploring an ontological theory derived from Bohmian mechanics' assumptions but with momentum as the ontological variable. In Bohmian theory, the choice of position as the ontological variable is neither self-evident nor obvious, and Bohm-like theories with a different choice of the ontological variable can also be constructed. I present an experimental comparison of Bohmian theory and a momentum-ontologic Bohmian-like theory and argue for the importance and need for making and justifying the choice of ontology.
The ability to gather information and to act upon it is at the core of every learning agent. But what is the impact of quantum mechanics on an agent's ability to sense external inputs and to translate them into actions? Here we address the question for a prototype task of learning agency at the quantum scale: rotating a single spin based on information gathered by a single atom. We determine the ultimate performance limit for this task, revealing a fundamental tradeoff between entanglement at the sensing stage and coherence at the action stage: if the single-atom sensor is not entangled with the quantum system serving as the agent's internal memory, then the best learning strategy requires a coherent transfer of quantum information from the sensor to the system that controls the agent's actions. In contrast, if the sensor is initially entangled with the agent's memory, then the transfer of quantum information is no longer necessary. Our results indicate that the quantum properties of the sensor radically affect the optimal way to convert external stimuli into actions, revealing a link between quantum sensing and the behavior of quantum agents.
In the general theory of quantum measurement, one associates a positive-semidefinite operator on a ddd-dimensional Hilbert space to each of the nnn possible outcomes of an arbitrary measurement. In the special case of a projective measurement, these operators are pairwise Hilbert-Schmidt orthogonal, but when n>dn>dn>d, orthogonality is restricted by positivity. This restriction allows us to more precisely state the quantum adage: information gain of a system is always accompanied by unavoidable disturbance. Specifically, we investigate three properties of a measurement with L\"uders rule updating: its disturbance, a measure of how the expected postmeasurement state deviates from the input; its measurement strength, a measure of the intrinsic information producing capacity of the measurement; and its orthogonality, a measure of the degree to which the measurement operators differ from an orthonormal set. These quantities satisfy an information-disturbance tradeoff relation that highlights the additional role played by orthogonality. Finally, we assess several classes of measurements on these grounds and identify symmetric informationally complete quantum measurements as the unique quantum analogs of a perfectly informative and nondisturbing classical ideal measurement.
Any physical process can be represented as a quantum channel mapping an initial state to a final state. Hence it can be characterized from the point of view of communication theory, i.e., in terms of its ability to transfer information. Quantum information provides a theoretical framework and the proper mathematical tools to accomplish this. In this context the notion of codes and communication capacities have been introduced by generalizing them from the classical Shannon theory of information transmission and error correction. The underlying assumption of this approach is to consider the channel not as acting on a single system, but on sequences of systems, which, when properly initialized allow one to overcome the noisy effects induced by the physical process under consideration. While most of the work produced so far has been focused on the case in which a given channel transformation acts identically and independently on the various elements of the sequence (memoryless configuration in jargon), correlated error models appear to be a more realistic way to approach the problem. A slightly different, yet conceptually related, notion of correlated errors applies to a single quantum system which evolves continuously in time under the influence of an external disturbance which acts on it in a non-Markovian fashion. This leads to the study of memory effects in quantum channels: a fertile ground where interesting novel phenomena emerge at the intersection of quantum information theory and other branches of physics. A survey is taken of the field of quantum channels theory while also embracing these specific and complex settings.
In reinforcement learning, an agent interacts sequentially with an environment to maximize a reward, receiving only partial, probabilistic feedback. This creates a fundamental exploration-exploitation trade-off: the agent must explore to learn the hidden dynamics while exploiting this knowledge to maximize its target objective. While extensively studied classically, applying this framework to quantum systems requires dealing with hidden quantum states that evolve via unknown dynamics. We formalize this problem via a framework where the environment maintains a hidden quantum memory evolving via unknown quantum channels, and the agent intervenes sequentially using quantum instruments. For this setting, we adapt an optimistic maximum-likelihood estimation algorithm. We extend the analysis to continuous action spaces, allowing us to model general positive operator-valued measures (POVMs). By controlling the propagation of estimation errors through quantum channels and instruments, we prove that the cumulative regret of our strategy scales as O~(K)\widetilde{\mathcal{O}}(\sqrt{K})O(K​) over KKK episodes. Furthermore, via a reduction to the multi-armed quantum bandit problem, we establish information-theoretic lower bounds demonstrating that this sublinear scaling is strictly optimal up to polylogarithmic factors. As a physical application, we consider state-agnostic work extraction. When extracting free energy from a sequence of non-i.i.d. quantum states correlated by a hidden memory, any lack of knowledge about the source leads to thermodynamic dissipation. In our setting, the mathematical regret exactly quantifies this cumulative dissipation. Using our adaptive algorithm, the agent uses past energy outcomes to improve its extraction protocol on the fly, achieving sublinear cumulative dissipation, and, consequently, an asymptotically zero dissipation rate.