The shape of MIP* = RE

There’s a famous parable about a group of blind men encountering an elephant for the very first time. The first blind man, who had his hand on the elephant’s side, said that it was like an enormous wall. The second blind man, wrapping his arms around the elephant’s leg, exclaimed that surely it was a gigantic tree trunk. The third, feeling the elephant’s tail, declared that it must be a thick rope. Vehement disagreement ensues, but after a while the blind men eventually come to realize that, while each person was partially correct, there is much more to the elephant than initially thought.

6-blind-men-hans-1024x6541-1

Last month, Zhengfeng, Anand, Thomas, John and I posted MIP* = RE to arXiv. The paper feels very much like the elephant of the fable — and not just because of the number of pages! To a computer scientist, the paper is ostensibly about the complexity of interactive proofs. To a quantum physicist, it is talking about mathematical models of quantum entanglement. To the mathematician, there is a claimed resolution to a long-standing problem in operator algebras. Like the blind men of the parable, each are feeling a small part of a new phenomenon. How do the wall, the tree trunk, and the rope all fit together?

I’ll try to trace the outline of the elephant: it starts with a mystery in quantum complexity theory, curves through the mathematical foundations of quantum mechanics, and arrives at a deep question about operator algebras.

The rope: The complexity of nonlocal games

In 2004, computer scientists Cleve, Hoyer, Toner, and Watrous were thinking about a funny thing called nonlocal games. A nonlocal game G involves three parties: two cooperating players named Alice and Bob, and someone called the verifier. The verifier samples a pair of random questions (x,y) and sends x to Alice (who responds with answer a), and y to Bob (who responds with answer b). The verifier then uses some function D(x,y,a,b) that tells her whether the players win, based on their questions and answers.

All three parties know the rules of the game before it starts, and Alice and Bob’s goal is to maximize their probability of winning the game. The players aren’t allowed to communicate with each other during the game, so it’s a nontrivial task for them to coordinate an optimal strategy (i.e., how they should individually respond to the verifier’s questions) before the game starts.

The most famous example of a nonlocal game is the CHSH game (which has made several appearances on this blog already): in this game, the verifier sends a uniformly random bit x to Alice (who responds with a bit a) and a uniformly random bit y to Bob (who responds with a bit b). The players win if a \oplus b = x \wedge y (in other words, the sum of their answer bits is equal to the product of the input bits modulo 2).

What is Alice’s and Bob’s maximum winning probability? Well, it depends on what type of strategy they use. If they use a strategy that can be modeled by classical physics, then their winning probability cannot exceed 75\% (we call this the classical value of CHSH). On the other hand, if they use a strategy based on quantum physics, Alice and Bob can do better by sharing two quantum bits (qubits) that are entangled. During the game each player measures their own qubit (where the measurement depends on their received question) to obtain answers that win the CHSH game with probability \cos^2(\pi/8) \approx .854\ldots (we call this the quantum value of CHSH). So even though the entangled qubits don’t allow Alice and Bob to communicate with each other, entanglement gives them a way to win with higher probability! In technical terms, their responses are more correlated than what is possible classically.

The CHSH game comes from physics, and was originally formulated not as a game involving Alice and Bob, but rather as an experiment involving two spatially separated devices to test whether stronger-than-classical correlations exist in nature. These experiments are known as Bell tests, named after John Bell. In 1964, he proved that correlations from quantum entanglement cannot be explained by any “local hidden variable theory” — in other words, a classical theory of physics.1 He then showed that a Bell test, like the CHSH game, gives a simple statistical test for the presence of nonlocal correlations between separated systems. Since the 1960s, numerous Bell tests have been conducted experimentally, and the verdict is clear: nature does not behave classically.

Cleve, Hoyer, Toner and Watrous noticed that nonlocal games/Bell tests can be viewed as a kind of multiprover interactive proof. In complexity theory, interactive proofs are protocols where some provers are trying to convince a verifier of a solution to a long, difficult computation, and the verifier is trying to efficiently determine if the solution is correct. In a Bell test, one can think of the provers as instead trying to convince the verifier of a physical statement: that they possess quantum entanglement.

With the computational lens trained firmly on nonlocal games, it then becomes natural to ask about their complexity. Specifically, what is the complexity of approximating the optimal winning probability in a given nonlocal game G? In complexity-speak, this is phrased as a question about characterizing the class MIP* (pronounced “M-I-P star”). This is also a well-motivated question for an experimentalist conducting Bell tests: at the very least, they’d want to determine if (a) quantum players can do better than classical players, and (b) what can the best possible quantum strategy achieve?

Studying this question in the case of classical players led to some of the most important results in complexity theory, such as MIP = NEXP and the PCP Theorem. Indeed, the PCP Theorem says that it is NP-hard to approximate the classical value of a nonlocal game (i.e. the maximum winning probability of classical players) to within constant additive accuracy (say \pm \frac{1}{10}). Thus, assuming that P is not equal to NP, we shouldn’t expect a polynomial-time algorithm for this. However it is easy to see that there is a “brute force” algorithm for this problem: by taking exponential time to enumerate over all possible deterministic player strategies, one can exactly compute the classical value of nonlocal games.

When considering games with entangled players, however, it’s not even clear if there’s a similar “brute force” algorithm that solves this in any amount of time — forget polynomial time; even if we allow ourselves exponential, doubly-exponential, Ackermann function amount of time, we still don’t know how to solve this quantum value approximation problem. The problem is that there is no known upper bound on the amount of entanglement that is needed for players to play a nonlocal game. For example, for a given game G, does an optimal quantum strategy require one qubit, ten qubits, or 10^{10^{10}} qubits of entanglement? Without any upper bound, a “brute force” algorithm wouldn’t know how big of a quantum strategy to search for — it would keep enumerating over bigger and bigger strategies in hopes of finding a better one.

Thus approximating the quantum value may not even be solvable in principle! But could it really be uncomputable? Perhaps we just haven’t found the right mathematical tool to give an upper bound on the dimension — maybe we just need to come up with some clever variant of, say, Johnson-Lindenstrauss or some other dimension reduction technique.2

In 2008, there was promising progress towards an algorithmic solution for this problem. Two papers [DLTW, NPA] (appearing on arXiv on the same day!) showed that an algorithm based on semidefinite programming can produce a sequence of numbers that converge to something called the commuting operator value of a nonlocal game.3 If one could show that the commuting operator value and the quantum value of a nonlocal game coincide, then this would yield an algorithm for solving this approximation problem!

Asking whether this commuting operator and quantum values are the same, however, immediately brings us to the precipice of some deep mysteries in mathematical physics and operator algebras, far removed from computer science and complexity theory. This takes us to the next part of the elephant.

The tree: mathematical foundations of locality

The mystery about the quantum value versus the commuting operator value of nonlocal games has to do with two different ways of modeling Alice and Bob in quantum mechanics. As I mentioned earlier, quantum physics predicts that the maximum winning probability in, say, the CHSH game when Alice and Bob share entanglement is approximately 85%. As with any physical theory, these predictions are made using some mathematical framework — formal rules for modeling physical experiments like the CHSH game.

In a typical quantum information theory textbook, players in the CHSH game are usually modelled in the following way: Alice’s device is described a state space \mathcal{H}_A (all the possible states the device could be in), a particular state |\psi_A\rangle from \mathcal{H}_A, and a set of measurement operators \mathcal{M}_A (operations that can be performed by the device). It’s not necessary to know what these things are formally; the important feature is that these three things are enough to make any prediction about Alice’s device — when treated in isolation, at least. Similarly, Bob’s device can be described using its own state space \mathcal{H}_B, state |\psi_B\rangle, and measurement operators \mathcal{M}_B.

In the CHSH game though, one wants to make predictions about Alice’s and Bob’s devices together. Here the textbooks say that Alice and Bob are jointly described by the tensor product formalism, which is a natural mathematical way of “putting separate spaces together”. Their state space is denoted by \mathcal{H}_A \otimes \mathcal{H}_B. The joint state |\psi_{AB}\rangle describing the devices comes from this tensor product space. When Alice and Bob independently make their local measurements, this is described by a measurement operator from the tensor product of operators from \mathcal{M}_A and \mathcal{M}_B. The strange correlations of quantum mechanics arise when their joint state |\psi_{AB}\rangle is entangled, i.e. it cannot be written as a well-defined state on Alice’s side combined with a well-defined state on Bob’s side (even though the state space itself is two independent spaces combined together!)

The tensor product model works well; it satisfies natural properties you’d want from the CHSH experiment, such as the constraint that Alice and Bob can’t instantaneously signal to each other. Furthermore, predictions made in this model match up very accurately with experimental results!

This is the not the whole story, though. The tensor product formalism works very well in non-relativistic quantum mechanics, where things move slowly and energies are low. To describe more extreme physical scenarios — like when particles are being smashed together at near-light speeds in the Large Hadron Collider — physicists turn to the more powerful quantum field theory. However, the notion of spatiotemporal separation in relativistic settings gets especially tricky. In particular, when trying to describe quantum mechanical systems, it is no longer evident how to assign Alice and Bob their own independent state spaces, and thus it’s not clear how to put relativistic Alice and Bob in the tensor product framework!

In quantum field theory, locality is instead described using the commuting operator model. Instead of assigning Alice and Bob their own individual state spaces and then tensoring them together to get a combined space, the commuting operator model stipulates that there is just a single monolithic space \mathcal{H} for both Alice and Bob. Their joint state is described using a vector |\psi\rangle from \mathcal{H}, and Alice and Bob’s measurement operators both act on \mathcal{H}. The constraint that they can’t communicate is captured by the fact that Alice’s measurement operators commute with Bob’s operators. In other words, the order in which the players perform their measurements on the system does not matter: Alice measuring before Bob, or Bob measuring before Alice, both yield the same statistical outcomes. Locality is enforced through commutativity.

The commuting operator framework contains the tensor product framework as a special case4, so it’s more general. Could the commuting operator model allow for correlations that can’t be captured by the tensor product model, even approximately56? This question is known as Tsirelson’s problem, named after the late mathematician Boris Tsirelson.

There is a simple but useful way to phrase this question using nonlocal games. What we call the “quantum value” of a nonlocal game G (denoted by \omega^* (G)) really refers to the supremum of success probabilities over tensor product strategies for Alice and Bob. If they use strategies from the more general commuting operator model, then we call their maximum success probability the commuting operator value of G (denoted by \omega^{co}(G)). Since tensor product strategies are a special case of commuting operator strategies, we have the relation \omega^* (G) \leq \omega^{co}(G) for all nonlocal games G.

Could there be a nonlocal game G whose tensor product value is different from its commuting operator value? With tongue-in-cheek: is there a game G that Alice and Bob could succeed at better if they were using quantum entanglement at near-light speeds? It is difficult to find even a plausible candidate game for which the quantum and commuting operator values may differ. The CHSH game, for example, has the same quantum and commuting operator value; this was proved by Tsirelson.

If the tensor product and the commuting operator models are the same (i.e., the “positive” resolution of Tsirelson’s problem), then as I mentioned earlier, this has unexpected ramifications: there would be an algorithm for approximating the quantum value of nonlocal games.

How does this algorithm work? It comes in two parts: a procedure to search from below, and one to search from above. The “search from below” algorithm computes a sequence of numbers \alpha_1,\alpha_2,\alpha_3,\ldots where \alpha_d is (approximately) the best winning probability when Alice and Bob use a d-qubit tensor product strategy. For fixed d, the number \alpha_d can be computed by enumerating over (a discretization of) the space of all possible d-qubit strategies. This takes a doubly-exponential amount of time in d — but at least this is still a finite time! This naive “brute force” algorithm will slowly plod along, computing a sequence of better and better winning probabilities. We’re guaranteed that in the limit as d goes to infinity, the sequence \{ \alpha_d\} converges to the quantum value \omega^* (G). Of course the issue is that the “search from below” procedure never knows how close it is to the true quantum value.

This is where the “search from above” comes in. This is an algorithm that computes a different sequence of numbers \beta_1,\beta_2,\beta_3,\ldots where each \beta_d is an upper bound on the commuting operator value \omega^{co}(G), and furthermore as d goes to infinity, \beta_d eventually converges to \omega^{co}(G). Furthermore, each \beta_d can be computed by a technique known as semidefinite optimization; this was shown by the two papers I mentioned.

Let’s put the pieces together. If the quantum and commuting operator values of a game G coincide (i.e. \omega^* (G) = \omega^{co}(G)), then we can run the “search from below” and “search from above” procedures in parallel, interleaving the computation of the \{\alpha_d\} and \{ \beta_d\}. Since both are guaranteed to converge to the quantum value, at some point the upper bound \beta_d will come within some \epsilon to the lower bound \alpha_d, and thus we would have homed in on (an approximation of) \omega^* (G). There we have it: an algorithm to approximate the quantum value of games.

All that remains to do, surely, is to solve Tsirelson’s problem in the affirmative (that commuting operator correlations can be approximated by tensor product correlations), and then we could put this pesky question about the quantum value to rest. Right?

The wall: Connes’ embedding problem

At the end of the 1920s, polymath extraordinaire John von Neumann formulated the first rigorous mathematical framework for the recently developed quantum mechanics. This framework, now familiar to physicists and quantum information theorists everywhere, posits that quantum states are vectors in a Hilbert space, and measurements are linear operators acting on those spaces. It didn’t take long for von Neumann to realize that there was a much deeper theory of operators on Hilbert spaces waiting to be discovered. With Francis Murray, in the 1930s he started to develop a theory of “rings of operators” — today these are called von Neumann algebras.

The theory of operator algebras has since flourished into a rich and beautiful area of mathematics. It remains inseparable from mathematical physics, but has established deep connections with subjects such as knot theory and group theory. One of the most important goals in operator algebras has been to provide a classification of von Neumann algebras. In their series of papers on the subject, Murray and von Neumann first showed that classifying von Neumann algebras reduces to understanding their factors, the atoms out of which all von Neumann algebras are built. Then, they showed that factors of von Neumann algebras come in one of three species: type I, type II, and type III. Type I factors were completely classified by Murray and von Neumann, and they made much progress on characterizing certain type II factors. However progress stalled until the 1970s, when Alain Connes provided a classification of type III factors (work for which he would later receive the Fields Medal). In the same 1976 classification paper, Connes makes a casual remark about something called type II_1 factors7:

We now construct an embedding of N into \mathcal{R}. Apparently such an embedding ought to exist for all II_1 factors.

This line, written in almost a throwaway manner, eventually came to be called “Connes’ embedding problem”: does every separable II_1 factor embed into an ultrapower of the hyperfinite II_1 factor? It seems that Connes surmises that it does (and thus this is also called “Connes’ embedding conjecture“). Since 1976, this problem has grown into a central question of operator algebras, with numerous equivalent formulations and consequences across mathematics.

In 2010, two papers (again appearing on the arXiv on the same day!) showed that the reach of Connes’ embedding conjecture extends back to the foundations of quantum mechanics. If Connes’ embedding problem has a positive answer (i.e. an embedding exists), then Tsirelson’s problem (i.e. whether commuting operator can be approximated by tensor product correlations) also has a positive answer! Later it was shown by Ozawa that Connes’ embedding problem is in fact equivalent to Tsirelson’s problem.

Remember that our approach to compute the value of nonlocal games hinged on obtaining a positive answer to Tsirelson’s problem. The sequence of papers [NPA, DLTW, Fritz, JNPPSW] together show that resolving — one way or another — whether this search-from-below, search-from-above algorithm works would essentially settle Connes’ embedding conjecture. What started as a funny question at the periphery of computer science and quantum information theory has morphed into an attack on one of the central problems in operator algebras.

MIP* = RE

We’ve now ended back where we started: the complexity of nonlocal games. Let’s take a step back and try to make sense of the elephant.

Even to a complexity theorist, “MIP* = RE” may appear esoteric. The complexity classes MIP* and RE refer to a bewildering grabbag of concepts: there’s Alice, Bob, Turing machines, verifiers, interactive proofs, quantum entanglement. What is the meaning of the equality of these two classes?

First, it says that the Halting problem has an interactive proof involving quantum entangled provers. In the Halting problem, you want to decide whether a Turing machine M, if you started running it, would eventually terminate with a well-defined answer, or if it would get stuck in an infinite loop. Alan Turing showed that this problem is undecidable: there is no algorithm that can solve this problem in general. Loosely speaking, the best thing you can do is to just flick on the power switch to M, and wait to see if it eventually stops. If M gets stuck in an infinite loop — well, you’re going to be waiting forever.

MIP* = RE shows with the help of all-powerful Alice and Bob, a time-limited verifier can run an interactive proof to “shortcut” the waiting. Given the Turing machine M‘s description (its “source code”), the verifier can efficiently compute a description of a nonlocal game G_M whose behavior reflects that of M. If M does eventually halt (which could happen after a million years), then there is a strategy for Alice and Bob that causes the verifier to accept with probability 1. In other words, \omega^* (G_M) = 1. If M gets stuck in an infinite loop, then no matter what strategy Alice and Bob use, the verifier always rejects with high probability, so \omega^* (G_M) is close to 0.

By playing this nonlocal game, the verifier can obtain statistical evidence that M is a Turing machine that eventually terminates. If the verifier plays G_M and the provers win, then the verifier should believe that it is likely that M halts. If they lose, then the verifier concludes there isn’t enough evidence that M halts8. The verifier never actually runs M in this game; she has offloaded the task to Alice and Bob, who we can assume are computational gods capable of performing million-year-long computations instantly. For them, the challenge is instead to convince the verifier that if she were to wait millions of years, she would witness the termination of M. Incredibly, the amount of work put in by the verifier in the interactive proof is independent of the time it takes for M to halt!

The fact that the Halting problem has an interactive proof seems borderline absurd: if the Halting problem is unsolvable, why should we expect it to be verifiable? Although complexity theory has taught us that there can be a large gap between the complexity of verification versus search, it has always been a difference of efficiency: if solutions to a problem can be efficiently verified, then solutions can also be found (albeit at drastically higher computational cost). MIP* = RE shows that, with quantum entanglement, there can be a chasm of computability between verifying solutions and finding them.

Now let’s turn to the non-complexity consequences of MIP* = RE. The fact that we can encode the Halting problem into nonlocal games also immediately tells us that there is no algorithm whatsoever to approximate the quantum value. Suppose there was an algorithm that could approximate \omega^* (G). Then, using the transformation from Turing machines to nonlocal games mentioned above, we could use this algorithm to solve the Halting problem, which is impossible.

Now the dominoes start to fall. This means that, in particular, the proposed “search-from-below”/”search-from-above” algorithm cannot succeed in approximating \omega^* (G). There must be a game G, then, for which the quantum value is different from the commuting operator value. But this implies Tsirelson’s problem has a negative answer, and therefore Connes’ embedding conjecture is false.

We’ve only sketched the barest of outlines of this elephant, and yet it is quite challenging to hold it in the mind’s eye all at once9. This story is intertwined with some of the most fundamental developments in the past century: modern quantum mechanics, operator algebras, and computability theory were birthed in the 1930s. Einstein, Podolsky and Rosen wrote their landmark paper questioning the nature of quantum entanglement in 1935, and John Bell discovered his famous test and inequality in 1964. Connes’ formulated his conjecture in the ’70s, Tsirelson made his contributions to the foundations of quantum mechanics in the ’80s, and about the same time computer scientists were inventing the theory of interactive proofs and probabilistically checkable proofs (PCPs).

We haven’t said anything about the proof of MIP* = RE yet (this may be the subject of future blog posts), but it is undeniably a product of complexity theory. The language of interactive proofs and Turing machines is not just convenient but necessary: at its heart MIP* = RE is the classical PCP Theorem, with the help of quantum entanglement, recursed to infinity.

What is going on in this proof? What parts of it are fundamental, and which parts are unnecessary? What is the core of it that relates to Connes’ embedding conjecture? Are there other consequences of this uncomputability result? These are questions to be explored in the coming days and months, and the answers we find will be fascinating.

Acknowledgments. Thanks to William Slofstra and Thomas Vidick for helpful feedback on this post.


  1. This is why quantum correlations are called “nonlocal”, and why we call the CHSH game a “nonlocal game”: it is a test for nonlocal behavior. 
  2. A reasonable hope would be that, for every nonlocal game G, there is a generic upper bound on the number of qubits needed to approximate the optimal quantum strategy (e.g., a game G with Q possible questions and A possible answers would require at most, say, 2^{O(Q \cdot A)} qubits to play optimally). 
  3. In those papers, they called it the field theoretic value
  4. The space \mathcal{H} can be broken down into the tensor product \mathcal{H}_A \otimes \mathcal{H}_B, and Alice’s measurements only act on the \mathcal{H}_A space and Bob’s measurements only act on the \mathcal{H}_B space. In this case, Alice’s measurements clearly commute with Bob’s. 
  5. In a breakthrough work in 2017, Slofstra showed that the tensor product framework is not exactly the same as the commuting operator framework; he shows that there is a nonlocal game G where players using commuting operator strategies can win with probability 1, but when they use a tensor-product strategy they can only win with probability strictly less than 1. However the perfect commuting operator strategy can be approximated by tensor-product strategies arbitrarily well, so the quantum values and the commuting operator values of G are the same. 
  6. The commuting operator model is motivated by attempts to develop a rigorous mathematical framework for quantum field theory from first principles (see, for example algebraic quantum field theory (AQFT)). In the “vanilla” version of AQFT, tensor product decompositions between casually independent systems do not exist a priori, but mathematical physicists often consider AQFTs augmented with an additional “split property”, which does imply tensor product decompositions. Thus in such AQFTs, Tsirelson’s problem has an affirmative answer. 
  7. Type II_1 is pronounced “type two one”. 
  8. This is not the same as evidence that M loops forever! 
  9. At least, speaking for myself. 

Sense, sensibility, and superconductors

Jonathan Monroe disagreed with his PhD supervisor—with respect. They needed to measure a superconducting qubit, a tiny circuit in which current can flow forever. The qubit emits light, which carries information about the qubit’s state. Jonathan and Kater intensify the light using an amplifier. They’d fabricated many amplifiers, but none had worked. Jonathan suggested changing their strategy—with a politeness to which Emily Post couldn’t have objected. Jonathan’s supervisor, Kater Murch, suggested repeating the protocol they’d performed many times.

“That’s the definition of insanity,” Kater admitted, “but I think experiment needs to involve some of that.”

I watched the exchange via Skype, with more interest than I’d have watched the Oscars with. Someday, I hope, I’ll be able to weigh in on such a debate, despite working as a theorist. Someday, I’ll have partnered with enough experimentalists to develop insight.

I’m partnering with Jonathan and Kater on an experiment that coauthors and I proposed in a paper blogged about here. The experiment centers on an uncertainty relation, an inequality of the sort immortalized by Werner Heisenberg in 1927. Uncertainty relations imply that, if you measure a quantum particle’s position, the particle’s momentum ceases to have a well-defined value. If you measure the momentum, the particle ceases to have a well-defined position. Our uncertainty relation involves weak measurements. Weakly measuring a particle’s position doesn’t disturb the momentum much and vice versa. We can interpret the uncertainty in information-processing terms, because we cast the inequality in terms of entropies. Entropies, described here, are functions that quantify how efficiently we can process information, such as by compressing data. Jonathan and Kater are checking our inequality, and exploring its implications, with a superconducting qubit.

With chip

I had too little experience to side with Jonathan or with Kater. So I watched, and I contemplated how their opinions would sound if expressed about theory. Do I try one strategy again and again, hoping to change my results without changing my approach? 

At the Perimeter Institute for Theoretical Physics, Masters students had to swallow half-a-year of course material in weeks. I questioned whether I’d ever understand some of the material. But some of that material resurfaced during my PhD. Again, I attended lectures about Einstein’s theory of general relativity. Again, I worked problems about observers in free-fall. Again, I calculated covariant derivatives. The material sank in. I decided never to question, again, whether I could understand a concept. I might not understand a concept today, or tomorrow, or next week. But if I dedicate enough time and effort, I chose to believe, I’ll learn.

My decision rested on experience and on classes, taught by educational psychologists, that I’d taken in college. I’d studied how brains change during learning and how breaks enhance the changes. Sense, I thought, underlay my decision—though expecting outcomes to change, while strategies remain static, sounds insane.

Old cover

Does sense underlie Kater’s suggestion, likened to insanity, to keep fabricating amplifiers as before? He’s expressed cynicism many times during our collaboration: Experiment needs to involve some insanity. The experiment probably won’t work for a long time. Plenty more things will likely break. 

Jonathan and I agree with him. Experiments have a reputation for breaking, and Kater has a reputation for knowing experiments. Yet Jonathan—with professionalism and politeness—remains optimistic that other methods will prevail, that we’ll meet our goals early. I hope that Jonathan remains optimistic, and I fancy that Kater hopes, too. He prophesies gloom with a quarter of a smile, and his record speaks against him: A few months ago, I met a theorist who’d collaborated with Kater years before. The theorist marveled at the speed with which Kater had operated. A theorist would propose an experiment, and boom—the proposal would work.

Sea monsters

Perhaps luck smiled upon the implementation. But luck dovetails with the sense that underlies Kater’s opinion: Experiments involve factors that you can’t control. Implement a protocol once, and it might fail because the temperature has risen too high. Implement the protocol again, and it might fail because a truck drove by your building, vibrating the tabletop. Implement the protocol again, and it might fail because you bumped into a knob. Implement the protocol a fourth time, and it might succeed. If you repeat a protocol many times, your environment might change, changing your results.

Sense underlies also Jonathan’s objections to Kater’s opinions. We boost our chances of succeeding if we keep trying. We derive energy to keep trying from creativity and optimism. So rebelling against our PhD supervisors’ sense is sensible. I wondered, watching the Skype conversation, whether Kater the student had objected to prophesies of doom as Jonathan did. Kater exudes the soberness of a tenured professor but the irreverence of a Californian who wears his hair slightly long and who tattooed his wedding band on. Science thrives on the soberness and the irreverence.

Green cover

Who won Jonathan and Kater’s argument? Both, I think. Last week, they reported having fabricated amplifiers that work. The lab followed a protocol similar to their old one, but with more conscientiousness. 

I’m looking forward to watching who wins the debate about how long the rest of the experiment takes. Either way, check out Jonathan’s talk about our experiment if you attend the American Physical Society’s March Meeting. Jonathan will speak on Thursday, March 5, at 12:03, in room 106. Also, keep an eye out for our paper—which will debut once Jonathan coaxes the amplifier into synching with his qubit.

An equation fit for a novel

Archana Kamal was hunting for an apartment in Cambridge, Massachusetts. She was moving MIT, to work as a postdoc in physics. The first apartment she toured had housed John Updike, during his undergraduate career at Harvard. No other apartment could compete; Archana signed the lease.

The apartment occupied the basement of a red-brick building covered in vines. The rooms spanned no more than 350 square feet. Yet her window opened onto the neighbors’ garden, whose leaves she tracked across the seasons. And Archana cohabited with history.

Apartment photos

She’s now studying the universe’s history, as an assistant professor of physics at the University of Massachusetts Lowell. The cosmic microwave background (CMB) pervades the universe. The CMB consists of electromagnetic radiation, or light. Light has particle-like properties and wavelike properties. The wavelike properties include wavelength, the distance between successive peaks. Long-wavelength light includes red light, infrared light, and radio waves. Short-wavelength light includes blue light, ultraviolet light, and X-rays. Light of one wavelength and light of another wavelength are said to belong to different modes.

Wavelength

Does the CMB have nonclassical properties, impossible to predict with classical physics but (perhaps) predictable with quantum theory? The CMB does according to the theory of inflation. According to the theory, during a short time interval after the Big Bang, the universe expanded very quickly: Spacetime stretched. Inflation explains features of our universe, though we don’t know what mechanism would have effected the expansion.

According to inflation, around the Big Bang time, all the light in the universe crowded together. The photons (particles of light) interacted, entangling (developing strong quantum correlations). Spacetime then expanded, and the photons separated. But they might retain entanglement.

Detecting that putative entanglement poses challenges. For instance, the particles that you’d need to measure could produce a signal too weak to observe. Cosmologists have been scratching their heads about how to observe nonclassicality in the CMB. One team—Nishant Agarwal at UMass Lowell and Sarah Shandera at Pennsylvania State University—turned to Archana for help.

A sky full of stars

Archana studies the theory of open quantum systems, quantum systems that interact with their environments. She thinks most about systems such as superconducting qubits, tiny circuits with which labs are building quantum computers. But the visible universe constitutes an open quantum system.

We can see only part of the universe—or, rather, only part of what we believe is the whole universe. Why? We can see only stuff that’s emitted light that has reached us, and light has had only so long to travel. But the visible universe interacts (we believe) with stuff we haven’t seen. For instance, according to the theory of inflation, that rapid expansion stretched some light modes’ wavelengths. Those wavelengths grew longer than the visible universe. We can’t see those modes’ peak-to-peak variations or otherwise observe the modes, often called “frozen.” But the frozen modes act as an environment that exchanges information and energy with the visible universe.

We describe an open quantum system’s evolution with a quantum master equation, which I blogged about four-and-a-half years ago. Archana and collaborators constructed a quantum master equation for the visible universe. The frozen modes, they found, retain memories of the visible universe. (Experts: the bath is non-Markovian.) Next, they need to solve the equation. Then, they’ll try to use their solution to identify quantum observables that could reveal nonclassicality in the CMB.

Frozen modes

Frozen modes

Archana’s project caught my fancy for two reasons. First, when I visited her in October, I was collaborating on a related project. My coauthors and I were concocting a scheme for detecting nonclassical correlations in many-particle systems by measuring large-scale properties. Our paper debuted last month. It might—with thought and a dash of craziness—be applied to detect nonclassicality in the CMB. Archana’s explanation improved my understanding of our scheme’s potential. 

Second, Archana and collaborators formulated a quantum master equation for the visible universe. A quantum master equation for the visible universe. The phrase sounded romantic to me.1 It merited a coauthor who’d seized on an apartment lived in by a Pulitzer Prize-winning novelist. 

Archana’s cosmology and Updike stories reminded me of one reason why I appreciate living in the Boston area: History envelops us here. Last month, while walking to a grocery, I found a sign that marks the building in which the poet e. e. cummings was born. My walking partner then generously tolerated a recitation of cummings’s “anyone lived in a pretty how town.” History enriches our lives—and some of it might contain entanglement.

 

1It might sound like gobbledygook to you, if I’ve botched my explanations of the terminology.

With thanks to Archana and the UMass Lowell Department of Physics and Applied Physics for their hospitality and seminar invitation.

The paper that begged for a theme song

A year ago, the “I’m a little teapot” song kept playing in my head.

I was finishing a collaboration with David Limmer, a theoretical chemist at the University of California Berkeley. David studies quantum and classical systems far from equilibrium, including how these systems exchange energy and information with their environments. Example systems include photoisomers.

A photoisomer is a molecular switch. These switches appear across nature and technologies. We have photoisomers in our eyes, and experimentalists have used photoisomers to boost solar-fuel storage. A photoisomer has two functional groups, or collections of bonded atoms, attached to a central axis. 

Photoisomer

Your average-Joe photoisomer spends much of its life in equilibrium, exchanging heat with room-temperature surroundings. The molecule has the shape above, called the cis configuration. Imagine shining a laser or sunlight on the photoisomer. The molecule can absorb a photon, or particle of light, gaining energy. The energized switch has the opportunity to switch: One chemical group can rotate downward. The molecule will occupy its trans configuration.

Switch

The molecule now has more energy than it had while equilibrium, albeit less energy than it had right after absorbing the photon. The molecule can remain in this condition for a decent amount of time. (Experts: The molecule occupies a metastable state.) That is, the molecule can store sunlight. For that reason, experimentalists at Harvard and MIT attached photoisomers to graphene nanotubules, improving the nanotubules’ storage of solar fuel.

Teapot 1

With what probability does a photoisomer switch upon absorbing a photon? This question has resisted easy answering, because photoisomers prove difficult to model: They’re small, quantum, and far from equilibrium. People have progressed by making assumptions, but such assumptions can lack justifications or violate physical principles. David wanted to derive a simple, general bound—of the sort in which thermodynamicists specialize—on a photoisomer’s switching probability.

He had a hunch as to how he could derive such a bound. I’ve blogged, many times, about thermodynamic resource theories. Thermodynamic resource theories are simple models, developed in quantum information theory, for exchanges of heat, particles, information, and more. These models involve few assumptions: the conservation of energy, quantum theory, and, to some extent, the existence of a large environment (Markovianity). With such a model, David suspected, he might derive his bound.

Teapot 2

I knew nothing about photoisomers when I met David, but I knew about thermodynamic resource theories. I’d contributed to their development, to the theorems that have piled up in the resource-theory corner of quantum information theory. Then, the corner had given me claustrophobia. Those theorems felt so formal, abstract, and idealized. Formal, abstract theory has drawn me ever since I started studying physics in college. But did resource theories model physical reality? Could they impact science beyond our corner of quantum information theory? Did resource theories matter?

I called for connecting thermodynamic resource theories to physical reality four years ago, in a paper that begins with an embarrassing story about me. Resource theorists began designing experiments whose results should agree with our theorems. Theorists also tried to improve the accuracy with which resource theories model experimentalists’ limitations. See David’s and my paper for a list of these achievements. They delighted me, as a step toward the broadening of resource theories’ usefulness. 

Like any first step, this step pointed toward opportunities. Experiments designed to test our theorems essentially test quantum mechanics. Scientists have tested quantum mechanics for decades; we needn’t test it much more. Such experimental proposals can push experimentalists to hone their abilities, but I hoped that the community could accomplish more. We should be able to apply resource theories to answer questions cultivated in other fields, such as condensed matter and chemistry. We should be useful to scientists outside our corner of quantum information.

Teapot 3

David’s idea lit me up like photons on a solar-fuel-storage device. He taught me about photoisomers, I taught him about resource theories, and we derived his bound. Our proof relies on the “second laws of thermodynamics.” These abstract resource-theory results generalize the second law of thermodynamics, which helps us understand why time flows in only one direction. We checked our bound against numerical simulations (experts: of Lindbladian evolution). Our bound is fairly tight if the photoisomer has a low probability of absorbing a photon, as in the Harvard-MIT experiment. 

Experts: We also quantified the photoisomer’s coherences relative to the energy eigenbasis. Coherences can’t boost the switching probability, we concluded. But, en route to this conclusion, we found that the molecule is a natural realization of a quantum clock. Our quantum-clock modeling extends to general dissipative Landau-Zener transitions, prevalent across condensed matter and chemistry.

Teapot 4

As I worked on our paper one day, a jingle unfolded in my head. I recognized the tune first: “I’m a little teapot.” I hadn’t sung that much since kindergarten, I realized. Lyrics suggested themselves: 

I’m a little isomer
with two hands.
Here is my cis pose;
here is my trans.

Stand me in the sunlight;
watch me spin.
I’ll keep solar
energy in!

The song lodged itself in my head for weeks. But if you have to pay an earworm to collaborate with David, do.

The quantum steampunker by Massachusetts Bay

Every spring, a portal opens between Waltham, Massachusetts and another universe. 

The other universe has a Watch City dual to Waltham, known for its watch factories. The cities throw a festival to which explorers, inventors, and tourists flock. Top hats, goggles, leather vests, bustles, and lace-up boots dot the crowds. You can find pet octopodes, human-machine hybrids, and devices for bending space and time. Steam powers everything.

Watch City

Watch City Steampunk Festival

So I learned thanks to Maxim Olshanyi, a professor of physics at the University of Massachusetts Boston. He hosted my colloquium, “Quantum steampunk: Quantum information meets thermodynamics,” earlier this month. Maxim, I discovered, has more steampunk experience than I. He digs up century-old designs for radios, builds the radios, and improves upon the designs. He exhibits his creations at the Watch City Steampunk Festival.

Maxim photo

Maxim Olshanyi

I never would have guessed that Maxim moonlights with steampunkers. But his hobby makes sense: Maxim has transformed our understanding of quantum integrability.

Integrability is to thermalization as Watch City is to Waltham. A bowl of baked beans thermalizes when taken outside in Boston in October: Heat dissipates into the air. After half-an-hour, large-scale properties bear little imprint of their initial conditions: The beans could have begun at 112ºF or 99º or 120º. Either way, the beans have cooled.

Integrable systems avoid thermalizing; more of their late-time properties reflect early times. Why? We can understand through an example, an integrable system whose particles don’t interact with each other (whose particles are noninteracting fermions). The dynamics conserve the particles’ momenta. Consider growing the system by adding particles. The number of conserved quantities grows as the system size. The conserved quantities retain memories of the initial conditions.

Imagine preparing an integrable system, analogously to preparing a bowl of baked beans, and letting it sit for a long time. Will the system equilibrate, or settle down to, a state predictable with a simple rule? We might expect not. Obeying the same simple rule would cause different integrable systems to come to resemble each other. Integrable systems seem unlikely to homogenize, since each system retains much information about its initial conditions.

Boston baked beans

Maxim and collaborators exploded this expectation. Integrable systems do relax to simple equilibrium states, which the physicists called the generalized Gibbs ensemble (GGE). Josiah Willard Gibbs cofounded statistical mechanics during the 1800s. He predicted the state to which nonintegrable systems, like baked beans in autumnal Boston, equilibrate. Gibbs’s theory governs classical systems, like baked beans, as does the GGE theory. But also quantum systems equilibrate to the GGE, and Gibbs’s conclusions translate into quantum theory with few adjustments. So I’ll explain in quantum terms.

Consider quantum baked beans that exchange heat with a temperature-T environment. Let \hat{H} denote the system’s Hamiltonian, which basically represents the beans’ energy. The beans equilibrate to a quantum Gibbs state, e^{ - \hat{H} / ( k_{\rm B} T ) } / Z. The k_{\rm B} denotes Boltzmann’s constant, a fundamental constant of nature. The partition function Z enables the quantum state to obey probability theory (normalizes the state).

Maxim and friends modeled their generalized Gibbs ensemble on the Gibbs state. Let \hat{I}_m denote a quantum integrable system’s m^{\rm th} conserved quantity. This system equilibrates to e^{ - \sum_m \lambda_m \hat{I}_m } / Z_{\rm GGE}. The Z_{\rm GGE} normalizes the state. The intensive parameters \lambda_m’s serve analogously to temperature and depend on the conserved quantities’ values. Maxim and friends predicted this state using information theory formalized by Ed Jaynes. Inventing the GGE, they unlocked a slew of predictions about integrable quantum systems. 

Olchanyi__radioboard_comp_2015_picture

A radio built by Maxim. According to him, “The invention was to replace a diode with a diode bridge, in a crystal radio, thus gaining a factor of two in the output power.”

I define quantum steampunk as the intersection of quantum theory, especially quantum information theory, with thermodynamics, and the application of this intersection across science. Maxim has used information theory to cofound a branch of quantum statistical mechanics. Little wonder that he exhibits homemade radios at the Watch City Steampunk Festival. He also holds a license to drive steam engines and used to have my postdoc position. I appreciate having older cousins to look up to. Here’s hoping that I become half the quantum steampunker that I found by Massachusetts Bay.

With thanks to Maxim and the rest of the University of Massachusetts Boston Department of Physics for their hospitality.

The next Watch City Steampunk Festival takes place on May 9, 2020. Contact me if you’d attend a quantum-steampunk meetup!

Quantum conflict resolution

If only my coauthors and I had quarreled.

I was working with Tony Bartolotta, a PhD student in theoretical physics at Caltech, and Jason Pollack, a postdoc in cosmology at the University of British Columbia. They acted as the souls of consideration. We missed out on dozens of opportunities to bicker—about the paper’s focus, who undertook which tasks, which journal to submit to, and more. Bickering would have spiced up the story behind our paper, because the paper concerns disagreement.

Quantum observables can disagree. Observables are measurable properties, such as position and momentum. Suppose that you’ve measured a quantum particle’s position and obtained an outcome x. If you measure the position immediately afterward, you’ll obtain x again. Suppose that, instead of measuring the position again, you measure the momentum. All the possible outcomes have equal probabilities of obtaining. You can’t predict the outcome.

The particle’s position can have a well-defined value, or the momentum can have a well-defined value, but the observables can’t have well-defined values simultaneously. Furthermore, if you measure the position, you randomize the outcome of a momentum measurement. Position and momentum disagree.

Tug-of-war

How should we quantify the disagreement of two quantum observables, \hat{A} and \hat{B}? The question splits physicists into two camps. Pure quantum information (QI) theorists use uncertainty relations, whereas condensed-matter and high-energy physicists prefer out-of-time-ordered correlators. Let’s meet the camps in turn.

Heisenberg intuited an uncertainty relation that Robertson formalized during the 1920s,

\Delta \hat{A} \, \Delta \hat{B} \geq \frac{1}{i \hbar} \langle [\hat{A}, \hat{B}] \rangle.

Imagine preparing a quantum state | \psi \rangle and measuring \hat{A}, then repeating this protocol in many trials. Each trial has some probability p_a of yielding the outcome a. Different trials will yield different a’s. We quantify the spread in a values with the standard deviation \Delta \hat{A} = \sqrt{ \langle \psi | \hat{A}^2 | \psi \rangle - \langle \psi | \hat{A} | \psi \rangle^2 }. We define \Delta \hat{B} analogously. \hbar denotes Planck’s constant, a number that characterizes our universe as the electron’s mass does. 

[\hat{A}, \hat{B}] denotes the observables’ commutator. The numbers that we use in daily life commute: 7 \times 5 = 5 \times 7. Quantum numbers, or operators, represent \hat{A} and \hat{B}. Operators don’t necessarily commute. The commutator [\hat{A}, \hat{B}] = \hat{A} \hat{B} - \hat{B} \hat{A} represents how little \hat{A} and \hat{B} resemble 7 and 5. 

Robertson’s uncertainty relation means, “If you can predict an \hat{A} measurement’s outcome precisely, you can’t predict a \hat{B} measurement’s outcome precisely, and vice versa. The uncertainties must multiply to at least some number. The number depends on how much \hat{A} fails to commute with \hat{B}.” The higher an uncertainty bound (the greater the inequality’s right-hand side), the more the operators disagree.

fistfight-cloud

Heisenberg and Robertson explored operator disagreement during the 1920s. They wouldn’t have seen eye to eye with today’s QI theorists. For instance, QI theorists consider how we can apply quantum phenomena, such as operator disagreement, to information processing. Information processing includes cryptography. Quantum cryptography benefits from operator disagreement: An eavesdropper must observe, or measure, a message. The eavesdropper’s measurement of one observable can “disturb” a disagreeing observable. The message’s sender and intended recipient can detect the disturbance and so detect the eavesdropper.

How efficiently can one perform an information-processing task? The answer usually depends on an entropy H, a property of quantum states and of probability distributions. Uncertainty relations cry out for recasting in terms of entropies. So QI theorists have devised entropic uncertainty relations, such as

H (\hat{A}) + H( \hat{B} ) \geq - \log c. \qquad (^*)

The entropy H( \hat{A} ) quantifies the difficulty of predicting the outcome a of an \hat{A} measurement. H( \hat{B} ) is defined analogously. c is called the overlap. It quantifies your ability to predict what happens if you prepare your system with a well-defined \hat{A} value, then measure \hat{B}. For further analysis, check out this paper. Entropic uncertainty relations have blossomed within QI theory over the past few years. 

Blossom

Pure QI theorists, we’ve seen, quantify operator disagreement with entropic uncertainty relations. Physicists at the intersection of condensed matter and high-energy physics prefer out-of-time-ordered correlators (OTOCs). I’ve blogged about OTOCs so many times, Quantum Frontiers regulars will be able to guess the next two paragraphs. 

Consider a quantum many-body system, such as a chain of qubits. Imagine poking one end of the system, such as by flipping the first qubit upside-down. Let the operator \hat{W} represent the poke. Suppose that the system evolves chaotically for a time t afterward, the qubits interacting. Information about the poke spreads through many-body entanglement, or scrambles.

Spin chain

Imagine measuring an observable \hat{V} of a few qubits far from the \hat{W} qubits. A little information about \hat{W} migrates into the \hat{V} qubits. But measuring \hat{V} reveals almost nothing about \hat{W}, because most of the information about \hat{W} has spread across the system. \hat{V} disagrees with \hat{W}, in a sense. Actually, \hat{V} disagrees with \hat{W}(t). The (t) represents the time evolution.

The OTOC’s smallness reflects how much \hat{W}(t) disagrees with \hat{V} at any instant t. At early times t \gtrsim 0, the operators agree, and the OTOC \approx 1. At late times, the operators disagree loads, and the OTOC \approx 0.

Dove

Different camps of physicists, we’ve seen, quantify operator disagreement with different measures: Today’s pure QI theorists use entropic uncertainty relations. Condensed-matter and high-energy physicists use OTOCs. Trust physicists to disagree about what “quantum operator disagreement” means.

I want peace on Earth. I conjectured, in 2016 or so, that one could reconcile the two notions of quantum operator disagreement. One must be able to prove an entropic uncertainty relation for scrambling, wouldn’t you think?

You might try substituting \hat{W}(t) for the \hat{A} in Ineq. {(^*)}, and \hat{V} for the \hat{B}. You’d expect the uncertainty bound to tighten—the inequality’s right-hand side to grow—when the system scrambles. Scrambling—the condensed-matter and high-energy-physics notion of disagreement—would coincide with a high uncertainty bound—the pure-QI-theory notion of disagreement. The two notions of operator disagreement would agree. But the bound I’ve described doesn’t reflect scrambling. Nor do similar bounds that I tried constructing. I banged my head against the problem for about a year.

Handshake

The sky brightened when Jason and Tony developed an interest in the conjecture. Their energy and conversation enabled us to prove an entropic uncertainty relation for scrambling, published this month.1 We tested the relation in computer simulations of a qubit chain. Our bound tightens when the system scrambles, as expected: The uncertainty relation reflects the same operator disagreement as the OTOC. We reconciled two notions of quantum operator disagreement.

As Quantum Frontiers regulars will anticipate, our uncertainty relation involves weak measurements and quasiprobability distributions: I’ve been studying their roles in scrambling over the past three years, with colleagues for whose collaborations I have the utmost gratitude. I’m grateful to have collaborated with Tony and Jason. Harmony helps when you’re tackling (quantum operator) disagreement—even if squabbling would spice up your paper’s backstory.

 

1Thanks to Communications Physics for publishing the paper. For pedagogical formatting, read the arXiv version. 

What distinguishes quantum thermodynamics from quantum statistical mechanics?

Yoram Alhassid asked the question at the end of my Yale Quantum Institute colloquium last February. I knew two facts about Yoram: (1) He belongs to Yale’s theoretical-physics faculty. (2) His PhD thesis’s title—“On the Information Theoretic Approach to Nuclear Reactions”—ranks among my three favorites.1 

Over the past few months, I’ve grown to know Yoram better. He had reason to ask about quantum statistical mechanics, because his research stands up to its ears in the field. If forced to synopsize quantum statistical mechanics in five words, I’d say, “study of many-particle quantum systems.” Examples include gases of ultracold atoms. If given another five words, I’d add, “Calculate and use partition functions.” A partition function is a measure of the number of states, or configurations, accessible to the system. Calculate a system’s partition function, and you can calculate the system’s average energy, the average number of particles in the system, how the system responds to magnetic fields, etc.

Line in the sand

My colloquium concerned quantum thermodynamics, which I’ve blogged about many times. So I should have been able to distinguish quantum thermodynamics from its neighbors. But the answer I gave Yoram didn’t satisfy me. I mulled over the exchange for a few weeks, then emailed Yoram a 502-word essay. The exercise grew my appreciation for the question and my understanding of my field. 

An adaptation of the email appears below. The adaptation should suit readers who’ve majored in physics, but don’t worry if you haven’t. Bits of what distinguishes quantum thermodynamics from quantum statistical mechanics should come across to everyone—as should, I hope, the value of question-and-answer sessions:

One distinction is a return to the operational approach of 19th-century thermodynamics. Thermodynamicists such as Sadi Carnot wanted to know how effectively engines could operate. Their practical questions led to fundamental insights, such as the Carnot bound on an engine’s efficiency. Similarly, quantum thermodynamicists often ask, “How can this state serve as a resource in thermodynamic tasks?” This approach helps us identify what distinguishes quantum theory from classical mechanics.

For example, quantum thermodynamicists found an advantage in charging batteries via nonlocal operations. Another example is the “MBL-mobile” that I designed with collaborators. Many-body localization (MBL), we found, can enhance an engine’s reliability and scalability. 

Asking, “How can this state serve as a resource?” leads quantum thermodynamicists to design quantum engines, ratchets, batteries, etc. We analyze how these devices can outperform classical analogues, identifying which aspects of quantum theory power the outperformance. This question and these tasks contrast with the questions and tasks of many non-quantum-thermodynamicists who use statistical mechanics. They often calculate response functions and (e.g., ground-state) properties of Hamiltonians.

These goals of characterizing what nonclassicality is and what it can achieve in thermodynamic contexts resemble upshots of quantum computing and cryptography. As a 21st-century quantum information scientist, I understand what makes quantum theory quantum partially by understanding which problems quantum computers can solve efficiently and classical computers can’t. Similarly, I understand what makes quantum theory quantum partially by understanding how much more work you can extract from a singlet \frac{1}{ \sqrt{2} } ( | 0 1 \rangle - |1 0 \rangle ) (a maximally entangled state of two qubits) than from a product state in which the reduced states have the same forms as in the singlet, \frac{1}{2} ( | 0 \rangle \langle 0 | + | 1 \rangle \langle 1 | ).

As quantum thermodynamics shares its operational approach with quantum information theory, quantum thermodynamicists use mathematical tools developed in quantum information theory. An example consists of generalized entropies. Entropies quantify the optimal efficiency with which we can perform information-processing and thermodynamic tasks, such as data compression and work extraction.

Most statistical-mechanics researchers use just the Shannon and von Neumann entropies, H_{\rm Sh} and H_{\rm vN}, and perhaps the occasional relative entropy. These entropies quantify optimal efficiencies in large-system limits, e.g., as the number of messages compressed approaches infinity and in the thermodynamic limit.

Other entropic quantities have been defined and explored over the past two decades, in quantum and classical information theory. These entropies quantify the optimal efficiencies with which tasks can be performed (i) if the number of systems processed or the number of trials is arbitrary, (ii) if the systems processed share correlations, (iii) in the presence of “quantum side information” (if the system being used as a resource is entangled with another system, to which an agent has access), or (iv) if you can tolerate some probability \varepsilon that you fail to accomplish your task. Instead of limiting ourselves to H_{\rm Sh} and H_{\rm vN}, we use also “\varepsilon-smoothed entropies,” Rényi divergences, hypothesis-testing entropies, conditional entropies, etc.

Another hallmark of quantum thermodynamics is results’ generality and simplicity. Thermodynamics characterizes a system with a few macroscopic observables, such as temperature, volume, and particle number. The simplicity of some quantum thermodynamics served a chemist collaborator and me, as explained in the introduction of https://arxiv.org/abs/1811.06551.

Yoram’s question reminded me of one reason why, as an undergrad, I adored studying physics in a liberal-arts college. I ate dinner and took walks with students majoring in economics, German studies, and Middle Eastern languages. They described their challenges, which I analyzed with the physics mindset that I was acquiring. We then compared our approaches. Encountering other disciplines’ perspectives helped me recognize what tools I was developing as a budding physicist. How can we know our corner of the world without stepping outside it and viewing it as part of a landscape?

Plane

1The title epitomizes clarity and simplicity. And I have trouble resisting anything advertised as “the information-theoretic approach to such-and-such.”