{"id":"5defbca4-4213-434b-a96a-01c9bcfbbf5a","arxiv_id":"2606.13323","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":7.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Introduces homogeneous progress model and proves that for Z = N(-δ,1) with μ ≤ e^δ the growth rate R_μ equals (log^{1+o(1)} μ / μ) R_1.","lead":"The paper introduces a homogeneous progress model for evolution strategies where each mutation yields a fitness change drawn from a fixed distribution Z independent of current fitness. It provides asymptotic bounds on the expected growth rate of the steady-state (μ+1)-ES via sandwiching modified processes.","discovery_kind":"unclear","skeptic_critique":{"model":"grok-4.3","headline":"No significant objection identified","rationale":"The reader's weakest_assumption correctly flags the model definition, but that definition is the paper's starting point rather than an unstated assumption whose failure would invalidate the theorem inside the model. No load-bearing gap in the analysis technique itself is detectable without the proofs; therefore the reader's UNVERDICTED stance is not altered by an additional concern.","tokens_in":1779,"tokens_out":278,"duration_ms":20428,"concrete_test":"Extract the definitions of the two modified processes and the sandwiching inequalities from the full manuscript; verify that their growth rates provably bound R_μ and that the asymptotic extraction yields exactly the stated log^{1+o(1)} μ / μ factor.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is a mathematical result on the growth rate R_μ within the explicitly defined homogeneous progress model (invariant Z independent of position). The abstract describes a sandwiching technique via modified processes to bound the overlapping-generation (μ+1)-ES. No internal inconsistency, hidden assumption in the stated regime (Z = N(-δ,1), μ ≤ e^δ), or gap in the high-level argument is visible from the given material. The modeling premise is definitional rather than a hidden premise that could falsify the derivation.","agreement_with_reader":"disagree"},"referee_report":{"model":"grok-4.3","summary":"The paper introduces the homogeneous progress model for Evolution Strategies, in which the relative fitness improvement of any mutation is drawn from an invariant distribution Z (e.g., a mean-shifted Gaussian) independent of current fitness or position. It analyzes the expected growth rate R_μ of the continuous steady-state (μ+1)-ES by constructing modified processes whose rates provably sandwich the original overlapping-generation process. For Z = N(-δ,1) and μ ≤ e^δ the paper claims the asymptotic R_μ = (log^{1+o(1)} μ / μ) R_1.","tokens_in":1886,"tokens_out":412,"duration_ms":25359,"significance":"If the sandwiching argument and resulting asymptotic hold, the work supplies a rigorous, parameter-free scaling law for progress rates in a simplified model that approximates generic landscapes far from the optimum. The sandwiching technique itself is a methodological contribution for handling dependencies induced by overlapping generations. The model is positioned as a tool for intractable problems such as hyperparameter tuning.","major_comments":[{"comment":"Abstract: the central claim rests on a 'rigorous sandwiching argument' that yields the stated asymptotic together with explicit error terms and a treatment of the continuous steady-state; none of these derivations appear in sufficient detail to allow verification of the result beyond the high-level description.","section":"Abstract"},{"comment":"The modeling premise that Z is strictly invariant (independent of current fitness value or position) is definitional for the homogeneous progress model, yet the manuscript does not discuss the regime in which this approximation remains accurate or how deviations would affect the derived scaling.","section":"Abstract"}],"minor_comments":[{"comment":"Notation: the symbol R_μ is introduced without an explicit equation defining it in terms of the steady-state distribution or expected improvement.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the careful review and constructive comments. We address each major point below and outline the revisions we intend to make to strengthen the manuscript.","responses":[{"response":"The full derivations of the sandwiching construction, the proof that the modified processes bound the original overlapping-generation process, the analysis of the continuous steady-state, and the explicit error terms supporting the asymptotic appear in Sections 3 and 4. The abstract follows the conventional high-level style. To improve verifiability we will add a concise proof outline to the introduction and ensure every error term is stated explicitly with its derivation reference.","revision_made":"partial","referee_comment":"[Abstract] Abstract: the central claim rests on a 'rigorous sandwiching argument' that yields the stated asymptotic together with explicit error terms and a treatment of the continuous steady-state; none of these derivations appear in sufficient detail to allow verification of the result beyond the high-level description."},{"response":"We agree that an explicit discussion of the model's validity regime is useful. In the revised version we will insert a dedicated paragraph (or short subsection) after the model definition that identifies the distance-from-optimum regime in which invariance of Z is a reasonable approximation on typical smooth landscapes and that briefly indicates how moderate deviations from invariance would be expected to affect the leading scaling term.","revision_made":"yes","referee_comment":"[Abstract] The modeling premise that Z is strictly invariant (independent of current fitness value or position) is definitional for the homogeneous progress model, yet the manuscript does not discuss the regime in which this approximation remains accurate or how deviations would affect the derived scaling."}],"tokens_in":1429,"tokens_out":331,"duration_ms":19420,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main takeaway is that in the homogeneous progress model, where every mutation's fitness improvement is drawn from the same distribution Z no matter where you are, the (μ+1)-ES has an expected growth rate that scales as (log to the power 1 plus little-o of μ, divided by μ) times the rate for one individual, provided μ is at most e to the δ for a Gaussian Z with mean shift -δ.\n\nWhat stands out is the new model itself, which strips away the landscape details to focus on steady progress far from the optimum, and the sandwiching method they use to bound the overlapping generations case. They create modified processes whose rates sandwich the true one, allowing them to get the asymptotic without dealing directly with the parent dependencies.\n\nThis is a reasonable way to make the overlapping (μ+1) case tractable, and it gives a clean result that differs from non-overlapping strategies. The derivation seems rigorous based on the high-level description.\n\nOne limitation is the core assumption that Z stays invariant; this is by design for the model but means it only approximates problems where progress statistics don't shift much. The o(1) in the exponent also leaves room for the bound to be loose in finite cases. No other major issues jump out.\n\nThe work is aimed at people analyzing evolution strategies theoretically, especially those interested in how population size affects progress rates in simplified settings. A reader focused on black-box optimization or ES theory would get something from it.\n\nI would recommend sending it for peer review; the technique and the result in this model are worth a closer look by referees.","headline":"The paper gives an asymptotic for (μ+1)-ES growth rate in a new model where fitness improvements are drawn from a fixed distribution Z.","tokens_in":2363,"tokens_out":405,"would_cite":false,"duration_ms":20504,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"In the homogeneous progress model the (μ+1)-ES grows at rate (log^{1+o(1)} μ / μ) times the single-parent rate for Gaussian fitness shifts when μ ≤ e^δ.","keywords":["evolution strategies","homogeneous progress model","runtime analysis","(μ+1)-ES","growth rate","Gaussian distribution","population size","steady-state analysis"],"falsifier":"Running an actual optimization problem and measuring whether the empirical distribution of fitness gains stays statistically constant across widely different fitness levels.","tokens_in":2682,"feed_emoji":"📈","tokens_out":779,"duration_ms":29040,"temperature":0.7,"pith_summary":"The paper defines a homogeneous progress model in which every mutation produces a fitness improvement drawn from a fixed distribution Z that never changes with current position. This model is applied to the steady-state (μ+1)-ES that keeps overlapping generations by selecting the best μ out of μ+1 candidates. The analysis proceeds by constructing two modified processes whose growth rates provably bracket the true rate, making the bounding processes tractable. For the concrete case Z equal to a normal distribution with mean minus δ, the expected growth rate R_μ satisfies R_μ equals (log to the power 1 plus little-o of μ divided by μ) times R_1 whenever μ is at most e to the power δ. The result supplies a precise population-size tradeoff for optimization regimes far from the global optimum.","feed_headline":"Larger populations slow (μ+1)-ES to (log μ / μ) times single rate","feed_subtitle":"When fitness gains are drawn from a fixed distribution, the (μ+1) strategy's overlapping generations produce this scaling for Gaussian shift","key_machinery":"Modified processes that sandwich the growth rate of the original (μ+1)-ES, allowing the true rate to be bounded between two more tractable quantities.","core_discovery":"The paper establishes that in the homogeneous progress model defined by an invariant improvement distribution Z, the expected growth rate R_μ of the continuous steady-state (μ+1)-ES can be bounded by constructing simpler modified processes that sandwich the original process. When Z equals the normal distribution N(-δ, 1) and μ is at most e^δ, this technique yields the exact asymptotic relation R_μ = (log^{1 + o(1)} μ / μ) R_1.","pith_inferences":["Similar sandwiching arguments might yield growth-rate bounds for other selection rules or non-Gaussian Z distributions.","The scaling suggests that moderate population sizes could be preferable in practice once the logarithmic factor is taken into account.","The homogeneous model supplies a simple baseline for predicting behavior in black-box settings such as hyperparameter search.","The result could be tested by simulating the (μ+1)-ES on artificial landscapes whose improvement statistics are forced to remain stationary."],"forward_implications":["The growth rate of the (μ+1)-ES is asymptotically smaller than the (1+1)-ES rate by the factor log^{1+o(1)} μ divided by μ.","The bound holds only while μ stays at most e^δ for a Gaussian shift of size δ.","The model is intended for regimes far from the global optimum where exact fitness functions are intractable.","The sandwiching technique handles the dependencies created by overlapping generations in plus-selection."],"fun_headline_variants":["(μ+1)-ES R_μ equals (log^{1+o(1)} μ / μ) times R_1","Progress model shows μ reduces (μ+1)-ES rate by log μ over μ","(μ+1)-ES overlapping gens slow growth to (log μ / μ) R_1","Analysis: (μ+1)-ES expected rate (log μ / μ) of single parent"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The relative fitness improvement produced by any mutation is always drawn from the same fixed distribution Z no matter where the search currently stands.","fun_headline_variants_meta":{"raw":{"variants":["(μ+1)-ES R_μ equals (log^{1+o(1)} μ / μ) times R_1","Progress model shows μ reduces (μ+1)-ES rate by log μ over μ","(μ+1)-ES overlapping gens slow growth to (log μ / μ) R_1","Analysis: (μ+1)-ES expected rate (log μ / μ) of single parent"]},"model":"grok-4.3","cost_usd":0.008082,"raw_usage":{"total_tokens":3722,"prompt_tokens":763,"num_sources_used":0,"completion_tokens":101,"cost_in_usd_ticks":80824500,"prompt_tokens_details":{"text_tokens":763,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":2858,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":763,"tokens_out":101,"duration_ms":18939,"temperature":1.0,"reasoning_tokens":2858,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-27T04:59:23.041610+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"Running an actual optimization problem and measuring whether the empirical distribution of fitness gains stays statistically constant across widely different fitness levels.","supporting_citations":[],"review_version":1}