We review with a tutorial scope the information theory foundations of quantum statistical physics. Only a small proportion of the variables that characterize a system at the microscopic scale can be controlled, for both practical and theoretical reasons, and a probabilistic description involving the observers is required. The criterion of maximum von Neumann entropy is then used for making reasonable inferences. It means that no spurious information is introduced besides the known data. Its outcomes can be given a direct justification based on the principle of indifference of Laplace. We introduce the concept of relevant entropy associated with some set of relevant variables; it characterizes the information that is missing at the microscopic level when only these variables are known. For equilibrium problems, the relevant variables are the conserved ones, and the Second Law is recovered as a second step of the inference process. For non-equilibrium problems, the increase of the relevant entropy expresses an irretrievable loss of information from the relevant variables towards the irrelevant ones. Two examples illustrate the flexibility of the choice of relevant variables and the multiplicity of the associated entropies: the thermodynamic entropy (satisfying the Clausius–Duhem inequality) and the Boltzmann entropy (satisfying the H-theorem). The identification of entropy with missing information is also supported by the paradox of Maxwell’s demon. Spin-echo experiments show that irreversibility itself is not an absolute concept: use of hidden information may overcome the arrow of time.
We review with a tutorial scope the information theory foundations of quantum statistical physics. Only a small proportion of the variables that characterize a system at the microscopic scale can be controlled, for both practical and theoretical reasons, and a probabilistic description involving the observers is required. The criterion of maximum von Neumann entropy is then used for making reasonable inferences. It means that no spurious information is introduced besides the known data. Its outcomes can be given a direct justification based on the principle of indifference of Laplace. We introduce the concept of relevant entropy associated with some set of relevant variables; it characterizes the information that is missing at the microscopic level when only these variables are known. For equilibrium problems, the relevant variables are the conserved ones, and the Second Law is recovered as a second step of the inference process. For non-equilibrium problems, the increase of the relevant entropy expresses an irretrievable loss of information from the relevant variables towards the irrelevant ones. Two examples illustrate the flexibility of the choice of relevant variables and the multiplicity of the associated entropies: the thermodynamic entropy (satisfying the Clausius–Duhem inequality) and the Boltzmann entropy (satisfying the H-theorem). The identification of entropy with missing information is also supported by the paradox of Maxwell’s demon. Spin-echo experiments show that irreversibility itself is not an absolute concept: use of hidden information may overcome the arrow of time.
This chapter introduces the thermodynamics of information, a concept that finds wide use in physics and astrophysics. The link between physics and information began in 1867, with James Clerk Maxwell's thought experiment of a creature capable of beating the second law of thermodynamics. Entropy can therefore be said to represent the missing information about a system that can't be accounted for in macroscopic descriptions; the larger the entropy, the more the missing information. When Maxwell introduced the demon, the distribution of speeds of the atoms of a gas at a fixed temperature had only recently been established . The demon dies an entropic death because it must discard information previously collected. One may reasonably assume that a measurement procedure is fundamentally associated with a certain definite average entropy production, and that this restores concordance with the second law. The physicist Leo Szilard undertook a careful examination of the role played by the involvement of an observer.
We now turn to the subject of thermodynamics, and to describing the process of thermalization of quantum systems. A quantum system that interacts with another system large enough to obey the usual rules of thermodynamics, is referred to as an open system. The thermal system with which it interacts is often referred to as a thermal bath , or simply as the environment. We will use these terms interchangeably. Modeling the behavior of open quantum systems has many applications. One of these is to model the noise that they are subjected to by their surroundings, noise that we will wish to control using measurement and feedback loops in Chapter 5. There are also two important connections between quantum measurements, information theory, and thermodynamics. The first of these is that the thermodynamic entropy of a macroscopic system can be identified with the information-theoretic entropy that characterizes the lack of knowledge of an observer about the microstates of the system. In turn, if an observer is allowed to make microscopic measurements on the system, then this modifies the thermodynamic rules regarding the conversion of heat to work. Such is the subject of Maxwell’s demon. Curiously it took more than a century to understand how Maxwell’s demon, essentially a microscopic measurement and feedback process, fits within the second law of thermodynamics. In another twist, it turns out that the process of erasing information requires an increase in the thermodynamic entropy of the environment, and this increase is directly related to the amount of information erased.
We review a general theory of thermodynamics of information processing. The background of this topic is the recently-developed nonequilibrium statistical mechanics and quantum (and classical) information theory. These theories are closely related to the modern technologies to manipulate and observe small systems; for example, macromolecules and colloidal particles in the classical regime, and quantum-optical systems and quantum dots in the quantum regime. First, we review a generalization of the second law of thermodynamics to the situations in which small thermodynamic systems are subject to quantum feedback control. The generalized second law is expressed in terms of an inequality that includes the term of information obtained by the measurement, as well as the thermodynamic quantities such as the free energy. This inequality leads to the fundamental upper bound of the work that can be extracted by a “Maxwell's demon”, which can be regarded as a feedback controller with a memory that stores measurement outcomes. Second, we review generalizations of the second law of thermodynamics to the measurement and information erasure processes of the memory of the demon that is a quantum system. The generalized second laws consist of inequalities that identify the lower bounds of the energy costs that are needed for the measurement and the information erasure. The inequality for the erasure leads to the celebrated Landauer's principle for a special case. Moreover, these inequalities enable us to reconcile Maxwell's demon with the second law of thermodynamics. In these inequalities, thermodynamic quantities and information contents are treated on an equal footing. In fact, the inequalities are model-independent, so that they can be applied to a broad class of information processing. Therefore, these inequalities can be called the second law of “information thermodynamics”.
Deriving the laws of thermodynamics from a microscopic picture is a central quest of statistical mechanics.This tutorial focuses on the derivation of the first and second law for isolated and open quantum systems far from equilibrium, where such foundational questions also become practically relevant for emergent nanotechnologies.The derivation is based on a microscopic definition of five essential quantities: internal energy, thermodynamic entropy, work, heat, and temperature.These definitions are shown to satisfy the phenomenological laws of nonequilibrium thermodynamics for a large class of states and processes.The consistency with previous results is demonstrated.The framework applies to multiple baths including particle transport and accounts for processes with, e.g., a changing temperature of the bath, which is determined microscopically.An integral fluctuation theorem for entropy production is satisfied.In summary, this tutorial introduces a consistent and versatile framework to understand and apply the laws of thermodynamics in the quantum regime and beyond.
Thermodynamic entropy, as defined by Clausius, characterizes macroscopic observations of a system based on phenomenological quantities such as temperature and heat. In contrast, information-theoretic entropy, introduced by Shannon, is a measure of uncertainty. In this Letter, we connect these two notions of entropy, using an axiomatic framework for thermodynamics [E. H. Lieb and J. Yngvason Proc. R. Soc. 469, 20130408 (2013)]. In particular, we obtain a direct relation between the Clausius entropy and the Shannon entropy, or its generalization to quantum systems, the von Neumann entropy. More generally, we find that entropy measures relevant in nonequilibrium thermodynamics correspond to entropies used in one-shot information theory.
(English) The modern understanding of physics is deeply linked with the concept of information. The revival of the study of quantum mechanics in the form of quantum information is just an example of a more general trend showing how the incorporation of ideas from information theory into the practice of physics is not simply a fertile opportunity to find new results, but it also offers a radically new understanding of what physics should be describing: emblematic of this paradigm shift is the transition from the infinite dimensional space associated to the wave functions appearing in the Schrödinger equation, to the extreme simplicity of the modern cornerstone of quantum mechanics, the qubit. A special place in this context is taken by thermodynamics: on the one hand, because it was one of the first branches of physics in which the role of information was explicitly recognised; on the other, as the formal correspondence between Shannon and Boltzmann entropy hints at a deep connection between the two. Ultimately, it almost feels like these two theories will end up coinciding, and one will speak about thermodynamics just as information theory with erasure. We are still far away from this claim, but the continuous appearance of information quantifiers in genuinely thermodynamics setting, especially when characterising the dissipation, cannot but corroborate this belief. The aim of this thesis is to move further in the identification of the two theories, by focusing on some aspects of information geometry and showing how these naturally apply to the study of thermodynamic transformations. In particular, the main object of interest is the family of quantum Fisher information metrics, thoroughly studied in the first part of the thesis. In there we prove a fact that motivates the interest in this quantity: despite the statistical setting in which it was originally formulated, it has such a deep dynamical nature that all physical evolutions can actually be defined just in terms of their behaviour with respect to the Fisher information. In the second part of the thesis, we connect this discussion to the field of thermodynamics. In this context, we show that the Fisher information metrics naturally emerge in the description of the dissipation in near-isothermal transformations, that is whenever the driving is slow enough for the system to be close to equilibrium during the whole protocol. This shows another example of what seems to be a general rule: the naturality with which structures developed in the context of statistical inference and information theory apply to the study of entropy production. (Català) La comprensió moderna de la física està profundament relacionada amb el concepte d'informació. El renaixement de l'estudi de la mecànica quàntica en forma d'informació quàntica és només un exemple d'una tendència més general que mostra com la incorporació d'idees de la teoria de la informació a la pràctica de la física no és simplement una oportunitat fèrtil per trobar nous resultats, sinó que també ofereix una comprensió radicalment nova del que hauria de descriure la física: emblemàtica d'aquest canvi de paradigma és la transició de l'espai dimensional infinit associat a les funcions d'ona que apareixen a l'equació de Schrödinger, a l'extrema simplicitat de la pedra angular moderna de la mecànica quàntica, la qubit. Un lloc especial en aquest context ocupa la termodinàmica: d'una banda, perquè va ser una de les primeres branques de la física en què es va reconèixer explícitament el paper de la informació; de l'altra, ja que la correspondència formal entre l'entropia de Shannon i Boltzmann deixa entreveure una connexió profunda entre tots dos. En última instància, gairebé sembla que aquestes dues teories acabaran coincidint, i es parlarà de termodinàmica igual que de teoria de la informació amb esborrat. Encara estem lluny d'aquesta afirmació, però l'aparició contínua de quantificadors d'informació en un entorn genuïnament termodinàmic, especialment quan es caracteritza la dissipació, no pot deixar de corroborar aquesta creença. L'objectiu d'aquesta tesi és avançar més en la identificació de les dues teories, centrant-se en alguns aspectes de la geometria de la informació i mostrant com aquestes s'apliquen de manera natural a l'estudi de les transformacions termodinàmiques. En particular, el principal objecte d'interès és la família de mètriques d'informació quàntiques de Fisher, estudiades a fons a la primera part de la tesi. Allà demostrem un fet que motiva l'interès per aquesta magnitud: malgrat l'entorn estadístic en què es va formular originalment, té una naturalesa dinàmica tan profunda que totes les evolucions físiques es poden definir en realitat només en termes del seu comportament respecte a la Informació de Fisher. En la segona part de la tesi, connectem aquesta discussió amb el camp de la termodinàmica. En aquest context, mostrem que les mètriques d'informació de Fisher sorgeixen de manera natural en la descripció de la dissipació en transformacions gairebé isotèrmiques, és a dir, sempre que la conducció és prou lenta perquè el sistema estigui a prop de l'equilibri durant tot el protocol. Això mostra un altre exemple del que sembla ser una regla general: la naturalitat amb què les estructures desenvolupades en el context de la inferència estadística i la teoria de la informació s'apliquen a l'estudi de la producció d'entropia. (Español) La comprensión moderna de la física está enlazada al concepto de información. El renacimiento en el interés por la mecánica cuántica causado por el nacimiento de la información cuántica es solo un ejemplo de una tendencia más general que muestra cómo la incorporación de ideas de la teoría de la información en la práctica de la física no es simplemente una oportunidad fértil para encontrar nuevos resultados, sino que también ofrece una comprensión radicalmente nueva de lo que la física debería describir: emblemático de este cambio de paradigma es la transición del espacio de dimensión infinita asociado a las funciones de onda que aparecen en la ecuación de Schrödinger, a la extrema simplicidad de la reciente piedra angular de la mecánica cuántica, el qubit. Un espacio especial en este contexto lo ocupa la termodinámica: por un lado, porque fue una de las primeras ramas de la física en la que se reconoció explícitamente el papel de la información; por el otro, por el hecho que la correspondencia formal entre la entropía de Shannon y la de Boltzmann sugiere una conexión profunda entre las dos. En última instancia, casi parece que estas dos teorías terminarán coincidiendo, y se hablará sobre termodinámica como la teoría de la información con pérdidas. Todavía estamos lejos de esta afirmación, pero la continua aparición de cuantificadores de información en contextos genuinamente termodinámico, especialmente al momento de caracterizar la disipación, no puede no corroborar esta creencia. El objetivo de esta tesis es avanzar en la identificación de las dos teorías, centrándose en algunos aspectos de la geometría de la informa- ción y mostrando cómo se aplican naturalmente al estudio de las transformaciones termodinámicas. En particular, el principal objeto de interés es la familia de métricas de información cuántica de Fisher, ampliamente estudiada en la primera parte de la tesis. Allí probamos un hecho que motiva el interés en esta cantidad: a pesar del contexto estadístico en el que se formuló originalmente, tiene una natura dinámica tan profunda que todas las evoluciones físicas pueden definirse solo en términos de sus comportamientos con respecto a la Información de Fisher. En la segunda parte de la tesis, conectamos esta discusión con el campo de la termodinámica. En este contexto, mostramos que las métricas de información de Fisher emergen naturalmente en la descripción de la disipación en transformaciones casi-isotérmicas, es decir, siempre que el desplazamiento sea lo suficientemente lento como para que el sistema esté cerca del equilibrio durante todo el protocolo. Esto muestra otro ejemplo de lo que parece ser una regla general: la naturalidad con la que estructuras desarrolladas en el contexto de la inferencia estadística y de la teoría de la información se aplican al estudio de la producción de entropía.