{"id":"2c924790-80a7-4b41-bf8c-60b4ba44a9bf","arxiv_id":"1908.10435","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A review organizing nearly many-body-localized systems around a common picture of localized degrees of freedom coupled to slow, nontrivial thermal baths.","lead":"This paper is a review of systems that are almost, but not exactly, many-body localized, meaning they relax so slowly that they look localized on practical timescales. It argues that a wide range of these systems can be described as localized regions coupled to small, slow, spatially sparse heat baths.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unified bath picture overreaches for quasiperiodic systems: hyperuniformity removes the rare thermal regions that seed avalanches, and Sec. 2.2 concedes no explicit construction of the needed baths.","rationale":"The Reader's weakest_assumption is exactly the load-bearing concern: the review assumes all apparent instabilities thermalize, and flags the numerical evidence as weak or nonexistent. My stress-test sharpens this for quasiperiodic systems, where the assumption conflicts with the review's own statement that hyperuniformity forbids Griffiths rare regions. The review does not close the gap; it offers a speculative, unconstructed 'spectator spin' bath that also implies a mobility edge. This directly undermines the abstract's universal theme of localized degrees coupled to a thermal bath. I nevertheless keep the verdict at CONDITIONAL (i.e., unchanged) because the paper is an expert review that repeatedly flags its own assumptions, and the quasiperiodic discussion is presented as an open problem rather than as established fact. The concrete test is the natural computational check: a quasiperiodic version of the avalanche toy model would directly test whether the mechanism that justifies the bath picture in random systems operates at all without rare regions.","tokens_in":70263,"tokens_out":4921,"duration_ms":60662,"concrete_test":"Run a quasiperiodic analogue of the avalanche toy model of Ref. [27]: couple a finite random-matrix bath to a quasiperiodic Aubry-André Anderson chain via H_int = g sum_i exp(-i/xi) sigma_i^x, and measure the maximum distance to which the bath hybridizes l-bits as xi is tuned through the critical value log(2)/2, for system sizes L = 20, 40, 80. Compare with the random-disorder case. If the avalanche threshold is absent, or the required xi_c diverges, in the quasiperiodic background, then rare-region avalanches do not seed quasiperiodic thermalization, and the unified bath picture cannot cover that class.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that nearly MBL systems can be understood as approximately localized degrees of freedom coupled to a sparse, slowly relaxing network of thermal regions. For random 1D systems this is supported by the avalanche argument of Sec. 1.2.3, where a rare thermal inclusion destabilizes the localized chain above a critical localization length. The load-bearing step is the extension of this picture to quasiperiodic systems, where Sec. 2.2.1 explicitly states that hyperuniformity forbids the usual Griffiths rare regions: a quasiperiodic Hamiltonian cannot contain arbitrarily large, anomalously thermal spatial patches. The review therefore needs another source for the 'thermal bath'. The candidates offered in Sec. 2.2.2 are state-dependent 'spectator spins' or rare configurations of the initial state, but the text concedes that 'no explicit construction of these baths exists' and that this scenario implies a many-body mobility edge, which the paper is otherwise inclined to reject. Sec. 2.2.3 further notes that whether avalanches even start in quasiperiodic systems in d>1 is unclear. If quasiperiodic MBL is instead stable, or if its transition is governed by typical-region physics rather than rare-region avalanches, then the 'common theme' asserted in the abstract and Sec. 1.1 is not universal; quasiperiodic systems would be an entire class outside the unified bath paradigm. The paper labels the underlying assumption 'admittedly speculative' in Sec. 1.1, but the abstract and the strongest claim present the bath picture as a finding rather than a conjecture. This is the single most load-bearing soft spot: the review's synthesis depends on it, and the review itself supplies evidence against it for quasiperiodic disorder.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript is a review of dynamics and transport in nearly many-body-localized (MBL) systems, organized around a unifying claim: such systems are best viewed as approximately localized degrees of freedom coupled to a sparse, slowly relaxing set of thermal regions, i.e., self-generated or externally imposed baths. After introducing the l-bit description and the avalanche instability in Sec. 1, the review covers disordered and quasiperiodic systems on the thermal side of the MBL transition, symmetry- and dimensionality-constrained systems, coupled MBL-bath and multi-component systems, driven and prethermal systems, common dynamical signatures, and available numerical/theoretical methods. The authors are unusually transparent: Sec. 1.1 explicitly labels the central organizing assumption as 'admittedly speculative' and notes that numerical evidence for the relevant instabilities is 'weak or nonexistent.' The review is best read as a strongly argued research program or perspective rather than as a collection of established results.","tokens_in":70525,"tokens_out":5427,"duration_ms":64413,"significance":"If the unifying bath picture is correct, the review would tie together several seemingly distinct phenomena—subdiffusive transport, slow entanglement growth, 1/f noise, and exponentially slow heating in driven systems—as manifestations of localized l-bits interacting with rare, sluggish thermal regions. The paper also provides a useful catalog of recent RG schemes, spectral-function arguments, and numerical methods, and it explicitly enumerates open questions. The authors deserve credit for stating their central assumption plainly and for not overclaiming numerical support. However, the significance is conditional: the abstract's 'common theme' is asserted more strongly than the body's own caveats allow, and the quasiperiodic case, which occupies a central place in the narrative, lacks even a suggested microscopic construction of the required baths.","major_comments":[{"comment":"The abstract states the 'common theme' as an organizing conclusion: 'A theme common to many of these problems is that they can be understood in terms of approximately localized degrees of freedom coupled to a heat bath (or baths) consisting of thermal degrees of freedom.' Yet Sec. 1.1 characterizes the same statement as an explicit assumption that is 'admittedly speculative' and notes that 'by and large the numerical evidence for these instabilities is weak or nonexistent.' Because every later estimate of thermalization rates inherits this assumption, the manuscript should clearly mark the unified bath picture as a conjecture in the abstract and introduction, and should state which subsequent quantitative claims (e.g., avalanche thresholds, Griffiths exponents, prethermal heating rates) are conditional on it. As written, the abstract's level of certainty exceeds what the body supports.","section":"Abstract and Sec. 1.1"},{"comment":"This is the load-bearing weak point for the claimed universality. The text correctly explains that quasiperiodic hyperuniformity forbids the usual Griffiths rare regions, and then, when proposing 'spectator spins' or state-dependent rare configurations as the thermal bath, concedes that 'no explicit construction of these baths exists.' It also notes that that scenario implies a many-body mobility edge, which the review elsewhere is inclined to reject. If quasiperiodic MBL is in fact stable, or if its transition is controlled by typical-region resonances, then the quasiperiodic class is an entire counterexample to the unified bath paradigm. The manuscript should either provide a concrete microscopic construction for the quasiperiodic bath or explicitly present the quasiperiodic bath idea as an open, speculative scenario with a proposed numerical or experimental discriminator (e.g., the presence or absence of a many-body mobility edge, or the scaling of transport exponents with system size). At minimum, the abstract and Sec. 1.1 should be qualified so that quasiperiodic systems are not swept into the 'common theme' without a caveat.","section":"Sec. 2.2.1 and Sec. 2.2.2"},{"comment":"The avalanche criterion and the critical localization length ξc = log 2/2 are derived by modeling the thermal region as a featureless random matrix even after it has absorbed many localized spins. The paper itself acknowledges that this 'clearly cannot be strictly true' and appeals to toy-model numerics for support. Since this estimate underlies the KT-type RG of Sec. 1.5 and the instability claims for higher-dimensional and long-range-interacting systems in Sec. 3, the authors should either quantify the expected corrections to the random-matrix ansatz (for example, the effect of slow modes in the absorbed block) or explicitly label the avalanche criterion as a heuristic whose numerical support is limited to the special coupling structure of Ref. [27]. The honesty of the caveat is appreciated, but a review article should not leave a load-bearing numerical estimate without a stated range of validity.","section":"Sec. 1.2.3 and Sec. 1.5"}],"minor_comments":[{"comment":"The manuscript contains numerous typographical errors that should be corrected in revision: 'reallty' in Sec. 3.1, 'noot' and 'Jorrdan' in the Sec. 3.2.3 discussion, 'pertrubation' near the SDRG description, 'terminii' in the Fig. 4 caption, and inconsistent accent rendering of 'Rényi.'","section":"Throughout"},{"comment":"Two citations are to unpublished or in-preparation work: 'A. Scardicchio et al., in preparation' in Sec. 2.1.1 and 'V. Khemani, unpublished' in the Sec. 2.2.2 footnote. These should be replaced by published preprints or removed, since a review article should not rest claims on unverifiable sources.","section":"Sec. 2.1.1 and Sec. 2.2.2 footnotes"},{"comment":"The notation for the spectral function is inconsistent: Eq. (8) uses f_T(ω), while the surrounding text defines f_E(ω). Please unify the notation and define the argument clearly.","section":"Sec. 1.4.3 and Eq. (8)"},{"comment":"The phrase '1 /A' appears with an unusual space; this is presumably the inverse drive amplitude 1/A. Please fix the typesetting throughout the paragraph.","section":"Sec. 1.4.4"},{"comment":"The term 'hyperuniform' is used in a central argument but is not defined in the text; a brief parenthetical definition or a glossary entry would make the section accessible to readers outside the localization community.","section":"Sec. 2.2.1"}],"recommendation":"major_revision","confidential_remarks":"This is a review article whose central scientific claim is explicitly self-labeled as speculative. That alone should not trigger rejection—transparency is a strength—but the abstract currently overstates the claim, and the quasiperiodic section is the weakest link in the otherwise coherent narrative. I would be comfortable with publication after the authors (i) qualify the abstract and introductory framing, (ii) either construct or clearly demarcate the quasiperiodic bath scenario, and (iii) add a short statement of the regime of validity of the random-matrix avalanche estimate. The paper also leans heavily on the authors' own prior work; in a review this is acceptable, but the quasiperiodic discussion would benefit from a more distanced presentation that weighs competing scenarios more evenly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis is a review, not a research paper, and you should approach it as such. The authors synthesize a broad swath of nearly-MBL literature into a single picture: localized degrees of freedom coupled to small, slow, sparse thermal baths. That framing is genuinely useful for teaching and for organizing the field. The review does it well—the sections on l-bit dynamics, Griffiths effects, the avalanche argument, and symmetry-constrained systems are careful and the reasoning is transparent. They also deserve credit for flagging where the numerical evidence is weak or nonexistent.\n\nThe soft spot is the load-bearing assumption, stated in Sec. 1.1: they assume all apparent instabilities thermalize unless there are specific reasons otherwise. They call this 'admittedly speculative.' That's honest. But the abstract and several summary statements present the bath picture as a common theme, which drifts toward a finding. The stress-test concern about quasiperiodic systems lands: Sec. 2.2.1 notes hyperuniformity forbids usual Griffiths rare regions, and Sec. 2.2.2 concedes no explicit construction of the needed baths exists. The review itself supplies evidence against the universality of the avalanche-bath story for quasiperiodic systems. This doesn't sink the review, but it means the unified perspective is an organizing hypothesis, not an established result.\n\nMinor issues: several references are to unpublished work ('in preparation', 'unpublished'), which is irritating in a review meant to be a resource. The organization is good, but a clearer separation of established results from the working assumption would improve it.\n\nBottom line: specialists in MBL and slow relaxation will get value from this; it's a good entry point for students. It deserves serious peer review—the field needs careful reviews like this. I'd engage with it, but I'd read it as a thoughtful map of a speculative landscape, not as a proof of the unified bath scenario.","headline":"A careful, honest review that maps nearly-MBL physics onto a speculative bath picture; the quasiperiodic case is the obvious soft spot.","tokens_in":71105,"tokens_out":1945,"would_cite":true,"duration_ms":22472,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This review argues that nearly MBL systems—too disordered to thermalize quickly but not strictly localized—are unified by a common mechanism: approximately conserved l-bits coupled to sparse, slowly relaxing thermal regions.","keywords":["many-body localization","nearly MBL systems","l-bits","avalanche instability","Griffiths effects","subdiffusion","slow thermalization","Floquet prethermalization"],"falsifier":"Compute the growth of a single thermal inclusion coupled to a quasiperiodic (rather than random) l-bit chain at sizes where the avalanche threshold should be crossed; if the inclusion remains localized and no thermal avalanche propagates, the review's blanket assumption that all apparent instabilities lead to thermalization is falsified.","tokens_in":70011,"feed_emoji":"⏳","tokens_out":6990,"duration_ms":75815,"temperature":0.7,"pith_summary":"This review argues that the many systems that are 'nearly' many-body localized—too disordered to thermalize quickly, but not meeting the stringent conditions for true localization—share a common skeleton: approximately localized degrees of freedom (l-bits) coupled to rare, slow regions that act as a heat bath. It claims that once this skeleton is assumed, subdiffusive transport, anomalously slow entanglement growth, 1/f noise, fragility of linear response, and exponentially slow heating in driven systems are consequences of the same mechanism, not separate accidents. The authors assemble evidence across four families: thermal systems near the transition, systems whose symmetry or dimensionality forbids localization, multi-component systems with mixed localized and thermal parts, and periodically driven systems. The payoff is a testable, unified account of why equilibration rates can be far slower than microscopic timescales in a broad class of experiments.","feed_headline":"Nearly localized systems relax slowly: l-bits feed a sluggish bath","feed_subtitle":"Subdiffusion, 1/f noise, slow entanglement, and slow heating all follow from l-bits coupled to rare sluggish regions.","key_machinery":"The central object is the l-bit (local integral of motion), a dressed, approximately conserved spin that exists in typical regions of a strongly disordered system; in true MBL the l-bits remain quasilocal forever, while in nearly MBL systems they are coupled to rare thermal regions. The load-bearing mechanism is the avalanche: a thermal inclusion of sufficient size absorbs neighboring l-bits one by one, growing and eventually thermalizing the whole system, at a rate set by comparing the matrix element to the thermal block's level spacing. The review uses this coupling of l-bits to sparse baths, formalized in strong-randomness renormalization schemes of the MBL transition, to organize the dynamics of all nearly MBL systems it treats.","core_discovery":"The paper's central claim is that strict MBL is a special case, and the more common phenomenon is 'nearly MBL' behavior in systems that eventually thermalize (or heat up) but only after extraordinarily long times. It proposes that these systems are described by an effective picture of approximately conserved local operators (l-bits, which are dressed spins that act as nearly conserved integrals of motion) coupled to sparse thermal inclusions. The thermal inclusions are not ordinary baths: they are small, slow, or spatially heterogeneous, which is why transport is anomalous, e.g., subdiffusion with a continuously varying exponent, entanglement growth as a sublinear power law rather than linear, and low-frequency noise approaching 1/f. The review explicitly assumes that every apparent instability of the localized phase leads to thermalization unless there are clear reasons to think otherwise, and on that assumption constructs a unified phenomenology that applies across disordered, quasiperiodic, multi-component, and driven systems.","pith_inferences":["If the avalanche picture is generic, strict MBL is a fragile one-dimensional, short-range phenomenon, and experiments on larger, higher-dimensional, or quasiperiodic systems should ultimately see thermalization on sufficiently long timescales; the review stops short of predicting those timescales for realistic experimental systems.","The unifying picture suggests that nearly MBL physics is controlled more by the density, size distribution, and spectral slowness of thermal regions than by the microscopic origin of disorder, which could be tested by comparing random and quasiperiodic systems with matched thermal-region statistics.","The explicit assumption that all apparent instabilities thermalize could be probed by measuring autocorrelation functions and eigenstate entanglement in quasiperiodic one-dimensional systems at sizes beyond current numerical reach; if a finite fraction of eigenstates persists with area-law entanglement there, the unified picture would require revision.","Because the 'bath' in the review is often itself far from thermal equilibrium, a more complete theory may need to assign different effective temperatures to different families of l-bits, an extension the review does not develop."],"forward_implications":["A system showing slow, MBL-like signatures on experimental timescales need not be truly MBL; the same signatures can arise from l-bits coupled to rare sluggish regions, so observing them does not by itself prove the existence of an MBL phase.","On the thermal side of the MBL transition, the review predicts a Griffiths phase with subdiffusion, continuously varying exponents, and 1/f noise, with the subdiffusive exponent vanishing as the transition is approached.","In the MBL phase itself, generic response functions should follow power laws in frequency whose exponents flow toward 1/f near the transition, and the limit of zero frequency and the limit of linear response should not commute.","Systems constrained to thermalize by continuous non-Abelian symmetry, dimensionality, or long-range interactions can still show MBL-like phenomenology for exponentially long times, and driven Floquet systems can heat to infinite temperature only exponentially slowly.","The MBL transition in random one-dimensional systems is controlled by a Kosterlitz-Thouless-like renormalization-group flow of thermal fraction and localization length, implying a diverging length scale and correspondingly slow finite-size scaling."],"supporting_citations":[{"why":"Provides the proof that a one-dimensional short-range random model has an MBL phase, establishing the baseline of true MBL from which nearly MBL systems deviate.","marker":"[24]"},{"why":"Introduces the rare-region avalanche instability of MBL by coupling l-bits to a random-matrix bath, the mechanism on which much of the review's unified picture depends.","marker":"[26]"},{"why":"Supplies toy-model numerics showing an avalanche transition as the localization length crosses a critical value, the main numerical support for the instability assumption.","marker":"[27]"},{"why":"Gives the Basko-Aleiner-Altshuler perturbative theory of MBL, defining the perturbative stability whose failure modes later sections enumerate.","marker":"[20]"},{"why":"Proposes a strong-randomness renormalization scheme for the MBL transition that yields the Griffiths/subdiffusive phenomenology used throughout the review.","marker":"[147]"},{"why":"Provides another strong-randomness RG scheme used to derive flow equations and low-frequency response behavior for the MBL phase.","marker":"[148]"},{"why":"Reports the initial numerical observation of subdiffusion, which the review interprets as a rare-region effect.","marker":"[156]"},{"why":"Develops the resonance-counting mechanism for low-frequency spectral functions in the MBL phase, central to the paper's account of 1/f noise and slow relaxation.","marker":"[141]"},{"why":"Presents the argument that continuous non-Abelian symmetries obstruct localization, which underlies the section on symmetry-constrained thermalization.","marker":"[196]"},{"why":"Discusses quasiperiodic MBL transitions and the role of avalanche spread, shaping the review's scenario for quasiperiodic systems.","marker":"[182]"}],"fun_headline_variants":["L-bits and slow baths: nearly MBL systems relax anomalously","Subdiffusion, 1/f noise, and slow heating from nearly MBL","Why some disordered systems thermalize only after eons","Dressed spins in rare regions: the nearly MBL story"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise, stated in Sec. 1.1, is that every apparent instability of the localized phase actually leads to thermalization unless there are specific theoretical reasons to believe otherwise—a premise the authors themselves call speculative, since numerical evidence for these instabilities is often weak or nonexistent.","fun_headline_variants_meta":{"raw":{"variants":["L-bits and slow baths: nearly MBL systems relax anomalously","Subdiffusion, 1/f noise, and slow heating from nearly MBL","Why some disordered systems thermalize only after eons","Dressed spins in rare regions: the nearly MBL story"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000238,"raw_usage":{"total_tokens":1573,"prompt_tokens":1069,"completion_tokens":504,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":685,"completion_tokens_details":{"reasoning_tokens":429}},"tokens_in":685,"tokens_out":504,"duration_ms":5416,"temperature":1.0,"reasoning_tokens":429,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:43:50.716518+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the growth of a single thermal inclusion coupled to a quasiperiodic (rather than random) l-bit chain at sizes where the avalanche threshold should be crossed; if the inclusion remains localized and no thermal avalanche propagates, the review's blanket assumption that all apparent instabilities lead to thermalization is falsified.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies toy-model numerics showing an avalanche transition as the localization length crosses a critical value, the main numerical support for the instability assumption."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proposes a strong-randomness renormalization scheme for the MBL transition that yields the Griffiths/subdiffusive phenomenology used throughout the review."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides another strong-randomness RG scheme used to derive flow equations and low-frequency response behavior for the MBL phase."}],"review_version":1}