Design boosters: from constant-time quantum chaos to ∞ \infty -designs and beyond
Soumik Ghosh, Arjun Mirani, Yihui Quek, Michelle Xu · 2025
Read Design boosters: from constant-time quantum chaos to ∞ \infty -designs and beyond by Soumik Ghosh, Arjun Mirani, Yihui Quek, Michelle Xu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-08543.txt.
Introduction
Soumik Ghosh and three co-authors ask whether measuring part of a quantum system can leave the rest more random than the whole. The effect, called deep thermalization, turns the unmeasured part into a set of states that mimics fully random states, known as a quantum state design.
The team proves a tight limit: starting from a global design of one strength, the leftover part can reach at most half that strength in general. They show the bound cannot be improved by building a case that hides a sign pattern in higher moments. Chaotic motion changes the picture, though. Under chaotic dynamics the leftover ensemble becomes fully random in the large-system limit, even from a weak global design.
Catalog record
| CREATOR | Soumik Ghosh, Arjun Mirani, Yihui Quek, Michelle Xu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 122 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-08543
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Design boosters: from constant-time quantum chaos to ∞ \infty -designs and beyond Soumik Ghosh, Arjun Mirani, Yihui Quek, Michelle Xu Abstract We study a counterintuitive property of ‘conditioning’ on the result of measuring a subsystem of a quantum state: such conditioning can boost design quality, at the cost of increased system size. We work in the setting of deep thermalization from many-body physics: starting from a bipartite state on a global system ( A , B ) (A,B) drawn from a k k -design, we measure system B B in the computational basis, keep the outcome and examine the state that remains in system A A , approximating the overall ensemble (the ‘projected ensemble’) by a k ′ k^{\prime} -design. We ask: how does the design quality change due to this procedure, or how does k ′ k^{\prime} compare to k k ? We give the first rigorous example of unitary dynamics generating a state such that, projection at very early (constant) times can boost design randomness. These dynamics are those of quantum chaos, modeled by the evolution of a Hamiltonian drawn from the Gaussian Unitary Ensemble (GUE). We show that, even though a state generated by such dynamics at constant time only forms a k = 𝒪 ( 1 ) k=\mathcal{O}(1) design, the projected ensemble is Haar-random (or a k ′ = ∞ k^{\prime}=\infty design) in the thermodynamic limit (i.e. when N B = ∞ N_{B}=\infty ). This phenomenon persists even with weaker and more physically realistic assumptions; our results can be appropriately applied to non-GUE Hamiltonians that nevertheless show likely chaotic signatures in their eigenbases. Moreover, we show that with no assumption on how the global state was generated, a k k -design experiences a degradation in design quality to k ′ = ⌊ k / 2 ⌋ k^{\prime}=\lfloor k/2\rfloor . This improves upon best prior results on the deep thermalization of designs. Together, our contributions argue for design boosting as a result of chaos and showcase a novel mechanism to generate good designs. I Introduction What happens to a random bipartite quantum state under projective measurement of one of its subsystems? Does the remaining subsystem get more or less random than the initial state, and does the answer depend on how the initial state was generated? We can phrase this as the question of how close the ensemble on the remaining subsystem is to a quantum state design. From a computer science perspective, this connects with the recent explosion of interest in generating quantum state and unitary designs in minimal circuit depth (see for instance BaHH 16 ; SHH 25 ; CSBH 25 ). The mechanism we outline above is an alternative technique for generating designs, aided by projected measurements. Measurements play a crucial role here: without measurements, a circuit with long-range two-qubit gates and with no limitation on the number of ancillas, must have depth at least \Omega(\log k+\log\log n/\varepsilon) to form an \epsilon -approximate state k -design CSBH 25 . When (non-adaptive) measurements are allowed, however, exact or approximate t -designs are achievable in constant circuit depth — thus bypassing the above lower bound — but with poly(n,t) many qubits HHB+ 20 ; TM 16 . These methods stem from the paradigm of measurement-based quantum computing (MBQC). They work by preparing a particular state known as a resource state, on which by doing a particular set of measurements, one can perform universal quantum computation on the unmeasured qubits. Our work strips away that structure, showing that by simply time-evolving with a random Hamiltonian from the Gaussian Unitary Ensemble and performing computational basis measurements on a subset of the qubits, the unmeasured qubits form a design. This all occurs in constant time, independent of the desired design quality; if we work in the thermodynamic limit (infinitely-many measured qubits), we can even sample Haar-randomly. From a physics perspective, our work builds on the paradigm of deep thermalization HC 22 ; CMH+ 23 as an instantiation of universal dynamics in many-body systems, the quest for which has animated much work in condensed matter physics for the past decade. Here, the term universality refers to the fact that the above procedure is agnostic to microscopic details: a random global Hamiltonian, and computational basis measurements, are all we need. Deep thermalization simply refers to the procedure of measuring a subsystem and looking at the state on the remaining subsystem. Our work exhibits improved parameter scalings as compared to previous work on deep thermalization, not only with respect to the number of qubits that must be measured but also in terms of the depth required to generate the initial state that is then measured. While MBQC experts will be no strangers to the fact that constant-depth circuits already suffice to generate designs, our work in the deep thermalization paradigm is the first to show that, constant-time Hamiltonian evolution already suffices. Let us now provide more motivation for studying the phenomenon of deep thermalization. Thermalization provides the broader context. The interplay between randomness, chaos, thermalization, and emergent universality is central to the foundations of statistical mechanics, both classical and quantum. In standard thermalization, we expect small subsystems of typical evolved states to look thermal Deu 91 ; Sre 94 ; RDO 08 ; NH 15 . Haar-random states have this property; so, up to moment order k , do state k-designs. Hence fast constructions of designs give fast proxies for thermal behavior MPSY 24 ; CHH+ 24 ; SHH 25 ; CSBH 25 ; FPVY 25 . 1 1 1 A quantum state k-design is an ensemble that matches the first k moments of the Haar ensemble. That said, global design quality \neq chaos. Recent low-depth constructions achieving strong global design guarantees can still miss hallmark chaotic diagnostics: for example, previously, it was believed that error bounds on approximate k -designs translated into information on k -point correlators in the system, and hence should be related to chaos; however, recent low-depth design constructions like SHH 25 do not have the correct out-of-time-order correlator or subsystem complexity behavior HS 25 to be chaotic. Global design quality is therefore an imperfect lens on chaos. This, along with inspiration from recent Rydberg-atom experiments, motivates a sharper diagnostic CSM+ 23 ; MSE+ 24 : prepare a state, projectively measure a large register B , and look at the residual state on A . We say an ensemble exhibits deep thermalization if, after this projection, the induced projected ensemble on A remains Haar-like — equivalently, if a global state k -design yields a projected state k^{\prime} -design for suitable k . Standard thermalization is the specific case of k^{\prime}=1 . This suggests the working conjecture: Globally chaotic systems should form excellent local projected ensemble designs. Now, since deep thermalization is a much stronger guarantee imposed on the local subsystem compared to thermalization, we ask whether this conjecture is equally universal as thermalization or whether it is a special property for certain tailored ensembles. We leave as an intriguing open question whether such deep thermalization objects could also have interesting applications to quantum computation. I.1 Our contributions In this work, as summarized below, we advance the web of connections between quantum designs, deep thermalization, and emergent universality,