{"id":"82062268-fb4a-4db5-9ab3-8a4e2be0904c","arxiv_id":"2505.01368","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A review of region identification in the method of regions, classifying regions into facet and hidden types and presenting recent all-loop results and conjectures.","lead":"This review explains how to find the correct set of regions in the method-of-regions technique for Feynman integrals, using Newton polytopes and a new split into facet and hidden regions. It summarizes all-loop results for wide-angle scattering and explores mode cascades in soft, collinear, mass, and Regge-limit expansions.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Hidden-region completeness rests on an unproven Landau-equation filter: the §2.3.2 recursion has no completeness proof, yet §3.2 and §5.2 use it to claim exhaustive four-loop enumerations and the Landshoff/Glauber picture.","rationale":"The reader's weakest assumption is the validity of the method of regions itself, which the paper explicitly lists as open question 1. That caveat is real but is a property of the whole framework and not something this review can be expected to settle. A more actionable, paper-specific gap lies in the hidden-region search. The central claim is that the facet/hidden classification plus polytope dissections gives a systematic way to identify all regions. For facet regions in the on-shell expansion, the paper points to a theorem proved in Ref. 17, so the main risk there is reliance on an external proof rather than an internal inconsistency. For hidden regions, the paper's own algorithm is only a necessary-condition filter, but the text extrapolates from that filter to exhaustive statements about three- and four-loop graphs: the ten graphs in Fig. 7 are said to be the complete set at three loops, and the 1081 four-loop graphs are used to conclude that all hidden-region graphs are extensions of Fig. 7. Even if the recursive deletion in Step 3 is correct for the examples shown, there is no proof that it terminates with a definitive yes/no in general. This matters because the Landshoff conjecture and the Regge-limit Glauber picture are motivated by exactly this enumeration. If the filter misses a Landau solution, the 'all hidden regions are Landshoff' conjecture would be false or at least unsupported. The reader's CONDITIONAL verdict remains appropriate; I would not change it, but the condition should explicitly require a completeness proof for the search algorithm or a clear statement that the hidden-region enumerations are only partial.","tokens_in":24291,"tokens_out":4907,"duration_ms":49794,"concrete_test":"Re-run the search algorithm of Section 2.3.2 on a complete enumeration of all 2-to-2 four-loop graphs, or on a finite class where the Landau equations can be solved independently by algebraic methods such as Gröbner bases or numerical homotopy continuation. Compare the set of graphs flagged as having 'potential pinch singularities' with the set of graphs admitting a positive real solution of F = 0 and α_e ∂F/∂α_e = 0 for all e. If any graph has a Landau solution but is not flagged, the algorithm is incomplete, and the four-loop 'all hidden regions' claim and the Landshoff conjecture lose their current support.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's systematic identification of hidden regions rests on the recursive Landau-equation search algorithm in Section 2.3.2. The algorithm is explicitly a necessary-condition filter: the 'potential pinch singularities' output only means no immediate obstruction was found, and Step 3 deletes variables for which one signed derivative vanishes identically and then iterates. No proof is given that this recursion is complete, i.e., that every positive solution of the Landau equations is detected and every 'no pinch' output is definitive. Yet Sections 3.2 and 5.2 use it to make strong enumerative claims: the ten three-loop graphs in Fig. 7 are presented as exhaustive for hidden regions, and at four loops the text states that the algorithm identifies 1081 such graphs and that 'all four-loop graphs with hidden regions can be constructed by adding one loop to one of the ten three-loop base graphs.' If the recursion has false negatives, hidden regions are missed, and the Landshoff/Glauber picture, which is the paper's main structural claim for hidden regions, is incomplete. The paper itself says the output is only a necessary condition, so there is an internal gap between the algorithm's stated capability and the all-loop conclusions drawn from it.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper is a review of systematic methods for identifying regions in the method-of-regions expansion of Feynman integrals. It introduces a classification of regions into facet regions (visible as lower facets of the Newton polytope of the Lee-Pomeransky polynomial) and hidden regions (interior to the polytope, arising from Landau-equation pinches). The proposed strategy for facet regions is graph-theoretic, based on minimum spanning (2-)trees, and for hidden regions it combines a recursive Landau-equation search with polytope dissections that convert hidden regions into facet regions of sub-polytopes. The framework is applied to the on-shell expansion for wide-angle scattering, where an all-loop theorem for facet regions is cited from Ref. 17 and hidden regions at three and four loops are claimed to be exhaustively enumerated (ten and 1081 graphs, respectively) with the conjecture that all hidden regions are Landshoff-scattering configurations. The soft, timelike-collinear, and heavy-to-light mass expansions are discussed, and the Regge-limit expansion for 2 to 2 forward scattering is treated, with claims of a cascade of momentum modes and hidden Glauber regions.","tokens_in":24607,"tokens_out":6251,"duration_ms":58320,"significance":"If the claims are correct, the paper provides a valuable unifying geometric framework for region identification, connecting Newton polytopes, Landau singularities, and graph topology, with potential applications to automating multiloop asymptotic expansions and to understanding factorization violations. The paper is clearly written, carefully distinguishes theorems from conjectures in most places, and includes instructive one-loop examples (Sudakov form factor, one-loop five-point graphs) that make the ideas concrete. The cited proofs in Refs. 16-18 provide a solid basis for the facet-region theorem and for the dissection methodology, and the paper explicitly labels its higher-loop statements (e.g., the Landshoff configuration conjecture) as conjectures. The significance is tempered, however, by the fact that the completeness of the hidden-region search algorithm is not established, and the Regge-limit results are stated without derivation or citation; as a result, the paper's central claim to provide a systematic way to identify all regions is stronger than the evidence presented.","major_comments":[{"comment":"The recursive search algorithm in §2.3.2 is explicitly a necessary-condition filter: the text states that the output 'represents only a necessary condition' and that cancellations might occur for unphysical (negative or complex) α_e values. Nevertheless, §3.2 uses the algorithm to make exhaustive statements—'At three-loop level, there are ten graphs with potential pinch singularities' (Fig. 7 caption: 'All the massless four-point three-loop graphs ...')—and §5.2 repeats the same enumeration for the Regge limit and further claims that 'all four-loop graphs with hidden regions can be constructed by adding one loop to one of the ten three-loop base graphs.' No completeness proof is given that the recursion detects every positive solution of the Landau equations and that every 'no pinch' output is definitive. Since false negatives would invalidate the Landshoff/Glauber picture as a complete classification of hidden regions, this is a load-bearing gap. The claims should be either backed by a proof (or a reference to one) or explicitly labeled as conjectures or partial results.","section":"§2.3.2, §3.2, §5.2"},{"comment":"The facet-region structure of the Regge-limit expansion is stated without any derivation or citation. In particular, the mode list in Eq. (23) (soft, soft·collinear, (collinear)^2 modes), Eq. (24) (higher-order analogues), the assertion that new modes emerge incrementally every two-loop order, and the statement that the upper/lower jet structure with Glauber modes is a facet region (Fig. 15) are all presented as established facts, but no reference to prior work or proof sketch is given. Unlike the on-shell-expansion theorem of §3.1, which is explicitly attributed to Ref. 17, these Regge-limit results appear to be new or unpublished. The reader cannot assess their validity; please provide references, derivations, or explicit conjecture labels.","section":"§5.2.1, Eqs. (23)–(24)"},{"comment":"The dissection strategy is described as producing 'a complete set of regions' after converting hidden regions to facet regions of sub-polytopes. Completeness requires that every hidden region of the original integral is captured by at least one sub-polytope and that the union of facet regions of the sub-polytopes does not introduce spurious regions. The paper refers to Ref. 18 for details, but the review's own use of this completeness to justify statements such as 'each of the ten graphs in Fig. 7 has a unique hidden region' goes beyond what is demonstrated here. Please state explicitly which parts of the dissection procedure are proven in Ref. 18 and which remain conjectural.","section":"§2.3.2"}],"minor_comments":[{"comment":"The characterization of Glauber scaling appears reversed: for kμ ∼ (λ, λ, λ^{1/2}), the two longitudinal (lightcone) components are O(λ) and the transverse components are O(λ^{1/2}), not vice versa as stated in the text.","section":"§5.1"},{"comment":"The paragraph beginning 'The soft expansion can be seen as a generalization of the on-shell expansion' is repeated almost verbatim in two consecutive paragraphs; the duplication should be removed.","section":"§4.1"},{"comment":"The phrase 'an search algorithm' should be 'a search algorithm'.","section":"§3.2"},{"comment":"The sentence 'Below let me list a few representative s' contains a typo; it should read 'a few representatives'.","section":"§6"}],"recommendation":"major_revision","confidential_remarks":"The paper leans heavily on the author's own previous work (Refs. 16–18), which is natural for a review of recent advances but means the central theorems are not independently checked in this manuscript. More importantly, the Regge-limit results in Section 5 appear to be unpublished new claims presented as review content; the editorial decision should weigh whether a review article is the appropriate venue for these results or whether they should be published with proofs elsewhere first. The completeness gap in the hidden-region enumeration, if not fixable, would weaken the paper's central claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Yao Ma's review is the clearest available summary of the facet-region/hidden-region picture for the method of regions. It is built mostly on Ma's own recent papers, but it presents that material with real pedagogical skill, and the speculative part about Landshoff/Glauber structure is honestly labeled. If you work on multiloop integrals or SCET, this is worth a careful read.\n\nCredit where it is due. The Newton-polytope setup is explained with concrete examples, the on-shell expansion theorem from Ref. 17 is stated accurately, and the one-loop Glauber examples make the notion of a hidden region tangible. The conjecture that all hidden regions in generic wide-angle scattering are Landshoff configurations is clearly marked as a conjecture, and the three-loop evidence is shown explicitly rather than waved at.\n\nThe main weakness is real. The search algorithm in Section 2.3.2 is a necessary-condition filter: it exits with \"potential pinch singularities\" when no immediate obstruction is found, and the recursion on F|αe=0 has no completeness proof. That is fine for a heuristic. But Section 3.2 then uses it to assert that the ten three-loop graphs are exhaustive, and Section 5.2 relies on the same algorithm to state that all four-loop graphs with hidden regions are obtained by adding one loop to those ten base graphs. The paper does not close the gap between \"no obstruction found\" and \"no hidden region missed.\" The Regge-limit mode cascade in Section 5.2.1 is also stated without derivation or citation; those modes may be right, but right now they look like claims to verify, not established results. Separately, a paragraph is duplicated across the two pages of Section 4.1, which a careful editor should have caught.\n\nThe paper openly notes that the method of regions itself lacks a rigorous proof, so the foundations are clearly flagged. The author does not oversell the all-loop conjectures.\n\nWho will get value: graduate students and practitioners who need to know which regions to include in an expansion, and developers of tools like Asy2 or AmpRed who want to see the geometric picture behind hidden regions. I would send it to peer review. The review would need revision: either prove completeness of the search algorithm or state the four-loop enumeration as conditional, provide support for the Regge-limit modes, and fix the duplicate paragraph.","headline":"Useful review of region identification, but the exhaustive hidden-region enumerations outrun the completeness of the search algorithm — worth a serious referee with revision.","tokens_in":25065,"tokens_out":4123,"would_cite":false,"duration_ms":41378,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.65.-w","04.62.+v"],"model":"deepseek-v4-flash","headline":"A Feynman integral's asymptotic expansion can be organized by splitting its regions into facet regions, visible as lower faces of a Newton polytope, and hidden regions, which require polytope dissection; in wide-angle scattering the…","keywords":["Feynman integrals","method of regions","asymptotic expansion","Newton polytope","hidden regions","Landshoff scattering","Glauber mode","Landau equations"],"falsifier":"Search the 1081 four-loop $2\\to 2$ wide-angle graphs identified as having potential pinch singularities: if any of them develops a hidden region whose momentum configuration is not a Landshoff scattering after polytope dissection, the conjecture that all hidden regions are Landshoff fails. A more basic test would be to find any Feynman integral in which the method-of-regions sum disagrees with a direct evaluation order by order in $\\lambda$, which would refute the foundational assumption on which every region list in this review depends.","tokens_in":24091,"feed_emoji":"🧮","tokens_out":5927,"duration_ms":56503,"temperature":0.7,"pith_summary":"This review argues that the correct list of regions in the method of regions—the terms in an asymptotic expansion of a Feynman integral—can be found systematically by splitting regions into two classes. Facet regions sit on the lower faces of a Newton polytope built from the Lee-Pomeransky polynomial and are described by simple power scalings of the integration parameters; hidden regions live inside the polytope and require extra scaling constraints. For on-shell wide-angle scattering the paper reports an all-loop theorem fixing all facet regions, and a conjecture, supported at three loops, that every hidden region is a Landshoff-scattering configuration. It also shows that the same graphs that hide Landshoff regions in wide-angle kinematics hide Glauber regions in the Regge limit, and that soft, collinear, and mass expansions each impose their own subgraph rules. The payoff is a route to enumerating all regions without constructing high-dimensional polytopes.","feed_headline":"Hidden Feynman regions all look like Landshoff scattering","feed_subtitle":"A polytope-based classification makes region-finding in asymptotic expansions a systematic all-loop procedure.","key_machinery":"The load-bearing object is the Newton polytope $\\Delta(P)$ of the Lee-Pomeransky polynomial $P(x;s)=U+F$, whose points encode the exponents of each monomial together with the scaling of its kinematic coefficient, $s\\cdot x^a\\mapsto(a,b)$. Lower facets of this polytope—facets whose inward normal has positive last entry—give region vectors whose first entries are the $\\lambda$-exponents of the parameters, so identifying facet regions reduces to finding lower facets. Hidden regions are handled by a second mechanism: Landau equations locate pinch singularities inside the integration domain, and a recursive search algorithm over $F_+$ and $F_-$ decides whether cancellations can occur; then a polytope dissection changes variables so that the pinch moves to an endpoint and the region becomes a facet of a sub-polytope. The graph-theoretic reformulation in terms of minimum spanning (2-)trees is what makes the all-loop facet theorem and the enumeration algorithms possible.","core_discovery":"The paper's central claim is that region identification in the method of regions is a solved problem for a broad class of asymptotic expansions once regions are split into two types. Facet regions are in one-to-one correspondence with the lower facets of the Newton polytope of the Lee-Pomeransky polynomial, so their parameter scalings are monomials $x_i\\sim\\lambda^{v_i}$; for the on-shell expansion of wide-angle scattering, an all-loop theorem asserts the complete configuration: one connected hard subgraph, one connected jet per external lightlike momentum, and a possibly disconnected soft subgraph, with constraints that exclude scaleless integrals. Hidden regions cannot be seen on polytope facets: they arise from cancellations in the $F$ polynomial satisfying Landau equations, and the paper's strategy is to dissect the polytope into sub-polytopes in which the pinch becomes an endpoint. Applied to $2\\to 2$ wide-angle scattering, this dissection yields exactly ten three-loop graphs with one hidden region each, and in every case the region is a Landshoff configuration where hard scatterings occur at distinct locations; at four loops all 1081 potential graphs contain three-loop subtopologies. The same ten graphs reappear as the hidden regions of the Regge-limit expansion, where the exchanged mode is Glauber rather than Landshoff.","pith_inferences":["If the Landshoff conjecture holds for generic wide-angle kinematics, then the minimal loop order at which hidden regions appear should track the number of distinct lightlike directions that enter separate hard vertices; the appearance of hidden regions already at two loops in $2\\to 3$ scattering is a first test of that counting.","Because the facet/hidden split is independent of numerator and spacetime dimension, the same dissection strategy should apply verbatim to phase-space integrals, where the paper notes only initial progress has been made.","A rigorous proof of the method of regions would presumably need to show that every possible pinch is either an endpoint (facet) or convertible to an endpoint by dissection; the classification in this review is a constructive candidate for that missing step.","The correspondence between hidden regions and disconnected off-shell subgraphs suggests that effective-theory descriptions built only from hard, collinear, and soft fields will miss the Landshoff and Glauber configurations that appear in individual integrals, even when amplitudes summed over regions remain correct."],"forward_implications":["For any wide-angle on-shell expansion, no momentum mode beyond hard, collinear, and soft can appear in a facet region; alternative scalings always give scaleless integrals.","Every hidden region of a $2\\to 2$ wide-angle scattering is a Landshoff configuration; at three loops there are exactly ten such graphs, and at four loops all candidate graphs contain one of these ten.","A region-finding algorithm can enumerate regions by graph structure alone, without building the polytope, at least for the on-shell expansion.","In the Regge limit the same hidden-region graphs exchange a Glauber momentum, so Glauber singularities and Landshoff singularities are two faces of the same pinch mechanism.","The soft, timelike-collinear, and heavy-to-light mass expansions each require extra jet compatibility constraints, and the mass and Regge expansions develop cascades of modes that appear every two loop orders."],"supporting_citations":[{"why":"Introduces the method of regions and the basic expansion formula $I=\\sum I(R_i)$ that this review organizes around.","marker":"[10]"},{"why":"Provides the Lee-Pomeransky representation on which the whole Newton-polytope construction rests.","marker":"[19]"},{"why":"Develops the graph-theoretical formulation of facet regions and the Landau-equation search algorithm that the review extends.","marker":"[16]"},{"why":"States and proves the on-shell-expansion region theorem giving all facet regions to all loops.","marker":"[17]"},{"why":"Presents the polytope-dissection strategy that converts hidden regions into facet regions and identifies the three-loop Landshoff graphs.","marker":"[18]"},{"why":"Gives the Landau equations used to detect the interior pinch singularities that define hidden regions.","marker":"[26]"},{"why":"Describes the Asy2 code's limitation with cancellations, motivating the need for dissection.","marker":"[21]"},{"why":"Supplies the Coleman-Norton interpretation connecting Landau solutions to the hard/collinear/soft spacetime picture.","marker":"[30]"}],"fun_headline_variants":["Every hidden Feynman region is a Landshoff configuration","Polytope dissection reveals all hidden regions are Landshoff","Hidden regions in asymptotic expansions: always Landshoff","Landshoff is the universal hidden Feynman region","Why hidden Feynman regions are all Landshoff"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole region list is only as trustworthy as the method of regions itself: the identity $I=\\sum I(R_i)$ with scaleless integrals set to zero has no rigorous proof or counterexample, and a failure of that identity would invalidate any region list built on it.","fun_headline_variants_meta":{"raw":{"variants":["Every hidden Feynman region is a Landshoff configuration","Polytope dissection reveals all hidden regions are Landshoff","Hidden regions in asymptotic expansions: always Landshoff","Landshoff is the universal hidden Feynman region","Why hidden Feynman regions are all Landshoff"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00024,"raw_usage":{"total_tokens":1489,"prompt_tokens":888,"completion_tokens":601,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":504,"completion_tokens_details":{"reasoning_tokens":529}},"tokens_in":504,"tokens_out":601,"duration_ms":5080,"temperature":1.0,"reasoning_tokens":529,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T04:19:35.278268+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search the 1081 four-loop $2\\to 2$ wide-angle graphs identified as having potential pinch singularities: if any of them develops a hidden region whose momentum configuration is not a Landshoff scattering after polytope dissection, the conjecture that all hidden regions are Landshoff fails. A more basic test would be to find any Feynman integral in which the method-of-regions sum disagrees with a direct evaluation order by order in $\\lambda$, which would refute the foundational assumption on which every region list in this review depends.","supporting_citations":[{"cited_title":"Landau, Nuclear Physics 13, 181 (1959)","cited_arxiv_id":null,"evidence_quote":"Gives the Landau equations used to detect the interior pinch singularities that define hidden regions."},{"cited_title":"Jantzen, A","cited_arxiv_id":null,"evidence_quote":"Describes the Asy2 code's limitation with cancellations, motivating the need for dissection."}],"review_version":1}