{"id":"d2732a4c-e352-4828-b31c-13909c934b58","arxiv_id":"2607.14740","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Under strong pinching and strong irreducibility of the linear parts, every line projection of a self-affine set attains the expected dimension; in the plane, strong irreducibility alone suffices.","lead":"For generic self-affine fractals, this paper proves that every one-dimensional projection has the expected Hausdorff dimension, and that the box-counting (Minkowski) dimension exists whenever its upper version is at most one. The results remove separation conditions and any restriction on the translations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified — the fixed-direction proof depends on Rapaport's entropy-dimension formula, not on Feng's exact-dimensionality theorem.","rationale":"The reader's verdict of ACCEPT is reasonable. The manuscript's proof is structurally sound: Theorem 1.1 is proved from Rapaport's entropy-dimension identity by local entropy averages, Theorem 1.2 is proved by extracting a strongly separated, dominated subsystem and applying the measure theorem, and Theorem 1.4 splits cleanly into proximal and non-proximal cases. I checked the internal steps of Proposition 3.1, Proposition 5.1, and Theorem 6.1 and found no obvious circularity or false inference. The reader's weakest-assumption analysis partly misses the mark: Feng's exact-dimensionality theorem appears only in Theorem 3.5, which is not used in the proof of Proposition 3.1 or later; the load-bearing external input is Rapaport's all-directions formula in Lemma 3.6. Thus the residual risk is the applicability of that published theorem, not Feng's. Since reliance on a cited published theorem with seemingly matching hypotheses is not, by itself, a flaw, I do not change the verdict. A verification of Rapaport's exact hypotheses would, however, remove the main residual uncertainty.","tokens_in":66407,"tokens_out":42795,"duration_ms":383379,"concrete_test":"Verify the statement of Rapaport [78, Theorem 1.12] against the hypotheses used in Lemma 3.6 and Theorem 3.2: exponential separation, fully supported Bernoulli ν, k-proximal and strongly k-irreducible for all k ∈ {1,...,d−1}, and the conclusion dime((proj_V)_*µ) = min{1,dim_L ν} for every V ∈ RP^{d−1}. If the theorem holds in this form, the proof is complete; if it has extra hypotheses (e.g., strong separation or a two-sided quasi-Bernoulli condition), check whether the strongly separated subsystem A'' and Bernoulli weights constructed in Theorem 6.1 satisfy them.","verdict_should_be":"UNCHANGED","load_bearing_attack":"No significant objection identified. The central chain (Theorem 1.1 and Theorem 6.1) is internally coherent: Proposition 3.1 uses Rapaport's all-directions entropy-dimension identity [78, Thm 1.12] in Lemma 3.6, not Feng's exact-dimensionality theorem; Theorem 3.5, where Feng enters, is not used afterward. The only residual risk is the usual one of relying on a heavy external theorem: if [78, Thm 1.12] had hypotheses beyond exponential separation, 1-proximality, and strong 1-irreducibility (e.g., a two-sided quasi-Bernoulli condition or strong separation), the fixed-direction lower bound in Proposition 3.1 would lose its starting point. The paper does not quote the full hypotheses of [78], so this deserves a check, but nothing in the manuscript indicates a mismatch. The self-identified gap in [6, Lemma 7.3] and its repair via ergodic decomposition are coherent.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves all-directions Marstrand–Mattila projection theorems for self-affine measures and sets in R^d under proximality/strong irreducibility hypotheses on the linear parts. Theorem 1.1 gives the expected dimension for projections of Bernoulli self-affine measures under exponential separation; Theorem 1.2 gives the set version under strong pinching, with no separation assumption; Theorem 1.4 is a sharp planar strengthening; Corollary 1.6 gives existence of the Minkowski dimension when the upper Minkowski dimension is at most one. The proof combines Rapaport's entropy-dimension formula and Feng's exact-dimensionality theorem with a local entropy averages argument, then passes from measures to sets via dominated and strongly separated subsystems. The paper also identifies and repairs a gap in the prior planar proof of [6, Lemma 7.3] by replacing an ergodicity claim with an ergodic decomposition argument.","tokens_in":66620,"tokens_out":24643,"duration_ms":244704,"significance":"If correct, the paper substantially advances the projection theory of self-affine sets: it removes the separation condition for sets, gives an all-directions result under hypotheses that are generic for tuples of matrices, proves a sharp planar statement, and gives a partial answer to the folklore Minkowski-dimension question. The paper is unusually careful: the repair of [6, Lemma 7.3] is explicit, the subsystem arguments are detailed, and Example 6.3 provides a concrete sharp counterexample showing that strong irreducibility cannot be weakened. The main theorems rely on heavy external results (Rapaport, Feng, Bougerol-Lacroix); the dependence is clearly identified and the checks given are plausible, though the exact hypotheses are not quoted.","major_comments":[],"minor_comments":[{"comment":"The set X_F as defined is empty: every c A_i^T with c≠0 is invertible and hence has full rank, not rank one. The set C = {c A_i^T : c∈R, i∈I*} is also not closed in general, so the compactness argument for K is invalid and the proof of spt(μ_F)=X_F collapses. This support identification is not used later in the paper, so the main theorems are unaffected, but the definition/proof should be corrected (e.g., define X_F via limits of normalized products, or delete the support claim).","section":"§3.1, Proposition 3.3"},{"comment":"Please state the precise hypotheses of Rapaport [78, Thm 1.12] and Feng [33, Thm 1.6(ii)] and verify them explicitly. The fixed-direction lower bound in Proposition 3.1 starts from Rapaport's all-directions identity; if that theorem has any unquoted hypothesis, the argument loses its starting point. The checks given in the paper (exponential separation, proximality, strong irreducibility, simple leading Lyapunov exponent) are plausible, but the current text does not allow the reader to verify them from the stated sources.","section":"§3.1, Theorem 3.5"},{"comment":"In the proof of ZF = ∪_i A_i ZF, the sentence beginning 'Writing i = ip ...' is garbled. Since V1 ∈ ZF and A_i V1 = V1 for every generator i, the inclusion is immediate; please rewrite this passage.","section":"§5.2, Lemma 5.3 proof"},{"comment":"The sentence 'As B1 likewise permutes the subspaces V1,...,Vm and preserves their dimensions, it fixes each Vj' is too strong: B1 could in principle permute the Vj in cycles. The intended contradiction only needs that B1 sends either admissible subspace (span{e3} or span{e1,e2}) to a subspace that is not admissible, which is true. Please adjust the wording.","section":"§4.3, Example 4.10(2)"},{"comment":"The notation dimM is used for both the upper Minkowski dimension (1.3) and the limit when it exists. Corollary 1.6 assumes 'dimM(X)≤1' using the upper dimension, then concludes that the Minkowski dimension exists. Please make this distinction explicit to avoid ambiguity.","section":"§1.2, Corollary 1.6"}],"recommendation":"minor_revision","confidential_remarks":"The stress-test concern about Feng's exact-dimensionality does not land: Proposition 3.1 and Lemma 3.6 use Rapaport's entropy-dimension identity, not Feng's theorem; Theorem 3.5, where Feng enters, is not used in the fixed-direction argument. The X_F issue in Proposition 3.3 is genuine but local and unused. The central chain is sound; I recommend minor revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Kolossváry and coauthors have a real result here: for a generic class of self-affine sets in R^d they prove the all-directions Marstrand–Mattila projection theorem, and for sets they remove the separation assumption entirely, not just for measures. The planar case is sharp, and the corollary on existence of Minkowski dimension under dim_M ≤ 1 is a genuine step on a folklore question. The paper also fixes a known gap in Bárány–Hochman–Rapaport by replacing an unproved ergodicity claim with an ergodic decomposition argument; that kind of honesty makes me trust the rest.\n\nWhat is new: Theorem 1.2 is the first all-directions result for self-affine sets in higher dimensions without separation and with arbitrary translations. The genericity theorem (1.3) is a nice byproduct. Example 6.3, showing irreducibility is not enough even under strong separation, is fully worked and convincing. The paper is unusually careful about its own dependencies, and the proof structure — passing through dominated and then strongly separated subsystems to apply the measure theorem — is coherent.\n\nSoft spots: the proof rests heavily on Rapaport's entropy-dimension identity [78, Thm 1.12], and the paper does not quote the full hypotheses of that theorem. The stress-test note says the actual route avoids Feng's theorem, and I agree after reading §§3–6: Feng enters only in Theorem 3.5, which is not used later. Still, a referee should check that the hypotheses of [78] match what is assumed here (exponential separation, proximality, strong irreducibility), because the fixed-direction lower bound in Proposition 3.1 depends on it. The paper is also long and very dense; I could not verify every line. That is a reason for careful refereeing, not for rejection.\n\nWho gets value: anyone working in fractal geometry, projection theory, and thermodynamic formalism. The Minkowski-dimension corollary will be cited widely. It deserves a serious referee.\n\nRecommendation: send to peer review. Yes, it is a major result. The referee should be someone comfortable with local entropy averages and multiplicative ergodic theorems, and should specifically check the use of Rapaport's theorem.","headline":"Genuinely new all-directions projection theorems for self-affine sets, with a rare combination of technical care and sharp examples; referee it, but verify the Rapaport input.","tokens_in":67168,"tokens_out":2592,"would_cite":true,"duration_ms":25539,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["28A80","28A75","37C45","37A05","15A75"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that under strong pinching and strong irreducibility of the linear parts, every line-projection of a self-affine set has the expected Hausdorff dimension, without any separation condition; in the plane, strong irreducibili","keywords":["self-affine sets","orthogonal projections","Hausdorff dimension","Minkowski dimension","local entropy averages","Zariski density","domination","Furstenberg measure"],"falsifier":"Compute the projection dimension of a concrete strongly pinching, strongly k-irreducible self-affine set in dimension three, for instance with algebraic matrix entries and generic translations; if any line projection falls strictly below min{1, dim_H X}, the set theorem fails. Alternatively, exhibit a Bernoulli self-affine measure meeting the stated hypotheses for which the all-directions entropy identity fails in some fixed direction.","tokens_in":66290,"feed_emoji":"📐","tokens_out":9227,"duration_ms":80332,"temperature":0.7,"pith_summary":"The paper shows that the random-direction conclusions of the classical projection theorem can be promoted to all-directions conclusions for self-affine measures and sets, provided the linear parts are sufficiently non-degenerate. For measures, exponential separation plus proximality and strong irreducibility in every exterior power force every one-dimensional projection to have the exact expected dimension. For sets, strong pinching in place of proximality does the same with no separation assumption, and in the plane strong irreducibility alone suffices. A corollary partially answers a long-standing folklore question by showing that self-affine sets of upper Minkowski dimension at most one have a Minkowski dimension. The interest lies in turning a generic direction statement into a universal one for a widely studied family of fractals.","feed_headline":"Self-affine sets project to the expected dimension on every line","feed_subtitle":"No line direction is exceptional for these fractals, and the Minkowski dimension exists when ≤ 1.","key_machinery":"The carrying mechanism is the local entropy averages method for the fixed-direction lower bound: a suspension flow over the symbolic space times projective space has a roof function recording the log contraction rate in the current direction, and a family of stopping-time cutsets converts typical-direction entropy identities (from an entropy-dimension formula and an exact-dimensionality theorem for the relevant self-affine measure) into a lower bound for every fixed direction. For the set theorem, the argument extracts a dominated and strongly separated subsystem, through a sandwiching construction with bounded prefix and suffix words and a packing argument, whose affinity dimension essentia","core_discovery":"The central discovery is an all-directions projection theorem: if the linear parts of an affine iterated function system are strongly pinching and strongly irreducible in every exterior power, then for every line through the origin the orthogonal projection of the self-affine set has Hausdorff dimension exactly min{1, dim_H X} and also equals min{1, dim_M X}; in particular, no exceptional directions exist. For self-affine measures, the same conclusion holds under the weaker proximality assumption together with exponential separation, giving exact-dimensionality of the projected measure in every direction. In the plane, the set theorem holds assuming only strong irreducibility, and an explici","pith_inferences":["If the entropy-dimension formula underlying the measure theorem is later established under weaker separation hypotheses, the same proof would immediately extend both the measure and set theorems.","The subsystem extraction technique (dominated plus strongly separated with nearly preserved affinity dimension) is likely reusable for other questions about self-affine sets, such as the full Minkowski dimension existence conjecture beyond the dim <= 1 case.","A natural testable extension would be to k-dimensional subspaces instead of lines; the methods appear adaptable once the local entropy averages machinery is developed on Grassmannians.","The sharp planar example suggests that in higher dimensions, irreducibility that is not strong will generally create exceptional directions, so the strong irreducibility assumption is probably essential rather than merely technical."],"forward_implications":["For every self-affine set satisfying the strong pinching and strong irreducibility hypotheses, the exceptional set of directions in the classical almost-everywhere projection theorem is empty.","If such a set has upper Minkowski dimension at most one, its Minkowski dimension exists and equals its Hausdorff dimension, a partial affirmative answer to the folklore question of whether Minkowski dimension always exists for self-affine sets.","In the plane, the exponential separation assumption used in earlier projection results can be removed entirely; strong irreducibility alone yields the all-directions conclusion.","The measure version gives that every one-dimensional projection of the self-affine measure is exact-dimensional with dimension min{1, dim_L(nu)}, where dim_L(nu) is the Lyapunov dimension.","The hypotheses hold for an open dense set of matrix tuples, so the result covers typical linear data."],"fun_headline_variants":["Every line projects self-affine sets to expected dimension","No exceptional directions for self-affine measures and sets","Minkowski dimension exists for self-affine sets with dim ≤ 1","Sharp projection result: planarity suffices for self-affine sets"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"Everything rests on the validity of an entropy-dimension formula and an exact-dimensionality theorem for the Bernoulli self-affine measure; if either theorem has hidden restrictions in this setting, the fixed-direction lower bound that starts the proof loses its starting point.","fun_headline_variants_meta":{"raw":{"variants":["Every line projects self-affine sets to expected dimension","No exceptional directions for self-affine measures and sets","Minkowski dimension exists for self-affine sets with dim ≤ 1","Sharp projection result: planarity suffices for self-affine sets"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000358,"raw_usage":{"total_tokens":1739,"prompt_tokens":668,"completion_tokens":1071,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":412,"completion_tokens_details":{"reasoning_tokens":1009}},"tokens_in":412,"tokens_out":1071,"duration_ms":8043,"temperature":1.0,"reasoning_tokens":1009,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T01:10:56.164869+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the projection dimension of a concrete strongly pinching, strongly k-irreducible self-affine set in dimension three, for instance with algebraic matrix entries and generic translations; if any line projection falls strictly below min{1, dim_H X}, the set theorem fails. Alternatively, exhibit a Bernoulli self-affine measure meeting the stated hypotheses for which the all-directions entropy identity fails in some fixed direction.","supporting_citations":[],"review_version":1}