{"id":"924e71b9-15c5-400a-adda-097b7ee7c167","arxiv_id":"2608.06051","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Finitely stable fractal dimensions of compact sets are always matched by some convergent sequence inside the set, and dimension-homogeneous sets admit one sequence matching many dimensions simultaneously.","lead":"This mathematics paper proves that common fractal dimensions can be 'witnessed' by a single convergent sequence of points inside a fractal. It shows when one sequence can match several dimensions at once, which sharpens understanding of fractal dimension definitions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Corollary 7.7 is not proved as stated: it needs continuity of s↦dim^s_B F, which is neither cited nor derived, so equality on a countable dense set of s-values does not automatically extend to all s.","rationale":"The paper’s central individual sequence-property theorems are convincingly argued: upper box dimension (Theorem 3.2), Assouad dimension and spectrum (Theorems 4.2 and 4.4), intermediate dimensions (Theorem 5.2), and dimension profiles (Theorem 6.4) all follow the announced strategy of locating a dimension point and extracting suitable finite sets at decreasing scales. The reader’s identification of the dimension-homogeneous/common-dimension-point assumption as load-bearing for the simultaneous results is fair, though for IFS attractors that assumption is automatically satisfied by considering sufficiently small cylinders, so it is not as restrictive as it first appears. However, the proof of Corollary 7.7 has a genuine gap: it silently passes from equality on countably many s-values to equality for all s, and the required continuity of the profile dimension is neither proved nor cited. This is precisely the kind of omitted support the review process should flag. The same section contains a smaller, repairable gap in Theorem 5.2, where a generic dense enumeration need not have finite initial segments converging in the Hausdorff metric; choosing an enumeration built from finite 1/k-nets fixes it. Neither issue undermines the main individual theorems, but Corollary 7.7 should be conditional on providing the missing continuity or on weakening the conclusion to a countable range of s.","tokens_in":14038,"tokens_out":41910,"duration_ms":417914,"concrete_test":"Determine analytically whether s↦dim^s_B F is continuous for every compact F⊂R^n. Start from the capacity definition (6.3) and try to prove |dim^t_B F − dim^s_B F|→0 as t→s, using the comparison C^s_r(F)≤C^t_r(F)(diam F/r)^{t−s} for t>s. If a full continuity proof can be supplied from the estimates of §6.1, add it (or a citation to [9,10]) to the proof of Corollary 7.7 and the concern is resolved. If the comparison yields only one-sided control and no continuity theorem exists in [9,10], test a concrete anisotropic example, e.g. E=[0,1]×C with C a Cantor set of dimension δ, computing dim^s_B E numerically on a fine grid of s near s=1+δ; a detected jump is a counterexample to Corollary 7.7 as written. Otherwise, restrict Corollary 7.7 to countably many s-values.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Corollary 7.7 claims that, for a dimension-homogeneous compact set E⊂R^n, a single sequence (a_i) satisfies dim^s_B(a_i)=dim^s_B E for all 0<s≤n. The proof is dismissed as “similar” to Corollary 7.6. That analogy is incomplete: Corollary 7.6 extends a countable dense set of θ-values to all θ by the known continuity of θ↦dim^θF and θ↦dim^θ_A F. Section 6 establishes the necessary properties for the profile dimension dim^s_B, namely monotonicity and finite stability, but it never states or proves continuity of s↦dim^s_B F for fixed F. Monotonicity alone does not suffice: two monotone functions can agree on a dense set and still differ at a jump. Lemma 7.4 and Corollary 7.5 can merge only countably many sequences, so the argument as written establishes equality only on a countable set of s-values. If there is any compact, even dimension-homogeneous, E for which s↦dim^s_B E is discontinuous, Corollary 7.7 fails in its present form and the simultaneous-profile claim in the abstract would need qualification. This is a missing proof or citation rather than a refutation; the individual sequence property for profiles, Theorem 6.4, appears sound.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces and studies the 'sequence property' for several fractal dimensions: a dimension has this property if every compact set of positive dimension contains a convergent sequence with the same dimension. The authors prove the sequence property for upper box dimension (Theorem 3.2), Assouad dimension (Theorem 4.2), the Assouad spectrum (Theorem 4.4), upper intermediate dimensions in locally compact spaces (Theorem 5.2), and box dimension profiles in R^n (Theorem 6.4). The main technical device is a localisation lemma (Lemma 2.5) showing that finitely stable dimensions always have a 'dimension point' x in every compact set, around which the set has full dimension at all scales. The paper also treats simultaneous representation by a single sequence: Corollaries 7.5 and 7.6 show that when several dimensions share a common dimension point, one sequence can witness all of them, provided the relevant functions of the parameter (θ or ϑ) are continuous. Corollary 7.7 makes a similar claim for dimension profiles, and Examples 7.1-7.3 show that without a common dimension point simultaneous representation can fail.","tokens_in":14331,"tokens_out":13909,"duration_ms":143921,"significance":"If the results are correct, they give a unified and largely constructive explanation of when a countable point set can have the same fractal dimension as a compact set, extending earlier work of Ivanov on upper box dimension. The localisation lemma and the merge lemma for sequences are clean and reusable tools, and the negative examples in Section 7 are valuable because they show that the 'common dimension point' condition is genuinely necessary. The use of intermediate pre-measure continuity (Lemma 5.1) and of capacity-based finite subsets (Lemma 6.3) gives the proofs a solid analytic grounding. The main caveat is that the simultaneous result for dimension profiles, Corollary 7.7, is not proved as written because it relies on an unstated continuity property in the profile parameter s.","major_comments":[{"comment":"The proof of Corollary 7.7 is not complete. It is dismissed as 'similar' to Corollary 7.6, but Corollary 7.6 crucially uses the known continuity of θ ↦ dim^θ F and ϑ ↦ dim^ϑ_A F to pass from equality on a countable dense set of parameters to equality for all parameters. For dimension profiles, Section 6 establishes monotonicity and finite stability of s ↦ dim^s_B F, but it neither states nor proves continuity of s ↦ dim^s_B F for a fixed compact set F, nor does it cite such a result. Consequently, applying Lemma 7.4 and Corollary 7.5 over a countable dense set of s-values yields equality only on that dense set. Monotonicity alone does not suffice: two monotone functions can agree on a dense set and differ at a jump. The statement of Corollary 7.7 therefore requires an additional continuity proof or a direct construction that avoids the dense-set argument; otherwise the claim is unproved as written.","section":"Corollary 7.7"},{"comment":"The use of the 'dimension homogeneous' hypothesis for dimension profiles is under-justified. The definition says every ball B(x,r) with x in E contains a bi-Lipschitz image of E, and the text then asserts that this makes every point a dimension point 'for a wide range of dimension definitions'. For Corollary 7.7 one needs specifically that dim^s_B(E∩B(x,r)) = dim^s_B E. This requires that dim^s_B is invariant under bi-Lipschitz embeddings and that a bi-Lipschitz copy of E contained in B(x,r) has the same profile as E. Section 6 defines dim^s_B via capacities and proves finite stability and the sequence property, but it does not state or prove bi-Lipschitz invariance of dim^s_B, nor is a reference given. A citation or a short proof is needed for this load-bearing step in Corollary 7.7.","section":"Section 7, after Corollary 7.5"}],"minor_comments":[{"comment":"The first sentence reads 'Let E⊂R^n be compact with dim^s_B = t > 0'; it should read 'with dim^s_B E = t > 0'.","section":"Section 6, Theorem 6.4 proof"},{"comment":"There is a garbled duplicated phrase in the proof: 'As in Example, 7.2, if (a_k)⊂E is a convergent sequence with a_k → x∈E then either a_k ∈ E_1 for all sufficiently large k so Let (a_k)_k ⊂E with a_k →x∈E. If x∈E_1...' The sentence needs to be rewritten.","section":"Example 7.3 proof"},{"comment":"The heading 'Propostition 6.2' contains a typo and should read 'Proposition 6.2'.","section":"Section 6.1"},{"comment":"The proof invokes Theorem 5.2 for θ_j in a countable dense subset of (0,1], but Theorem 5.2 is stated only for θ∈(0,1). The endpoint θ=1 is covered by Theorem 3.2, and the continuity cited from [12] is on (0,1]; this endpoint issue should be stated explicitly so the reader knows how θ=1 is included.","section":"Corollary 7.6(i)"}],"recommendation":"major_revision","confidential_remarks":"The core sequence-property theorems (Sections 3-6) appear sound and well motivated. The only substantive obstacle is Corollary 7.7, whose current proof lacks the needed continuity of s↦dim^s_B F; this is a local gap that a reference or a short additional argument could fix. I would not reject on this basis."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know: this paper answers a natural structural question for finitely stable fractal dimensions, and the core results are solid. It proves that upper box, Assouad, Assouad spectrum, upper intermediate dimensions, and box dimension profiles each have the sequence property: every compact set contains a convergent sequence whose dimension equals the set's dimension. The upper box case was known (Ivanov, cited as [18]); the genuinely new cases are Assouad, intermediate, and profiles, plus the negative examples showing you cannot always witness two dimensions at once without a common dimension point.\n\nWhat is good. The method is coherent. Lemma 2.5 (dimension points exist for any finitely stable dimension) is a nice localisation tool. Lemma 5.1, continuity of intermediate pre-measures in the Hausdorff metric, is a solid technical contribution that makes the intermediate dimension argument work, and the application to locally compact spaces is careful. The capacity argument for profiles (Section 6) is standard but clean. The negative examples 7.1-7.3 are instructive and correctly identify why simultaneous witnessing fails without a common dimension point. The paper is honest about prior attributions, flagging Ivanov's box dimension result explicitly.\n\nSoft spots. The main issue is Corollary 7.7. It claims that, for a dimension-homogeneous compact E, a single sequence witnesses dim^s_B E for all 0 < s <= n. The proof is dismissed as \"similar\" to Corollary 7.6. That analogy is incomplete. Corollary 7.6 works because theta-maps dim^theta E and the Assouad spectrum are known to be continuous in theta, so equality on a countable dense set of theta-values extends to all theta. For dimension profiles, the authors establish monotonicity and finite stability but do not establish continuity of s -> dim^s_B F, and monotonicity alone does not let you pass from a dense set to all s. As written, Corollary 7.7 only proves equality on a countable dense set of s-values. This is a missing argument or a missing citation, not a refutation: if continuity of profiles is known (likely from the capacity theory), it should be stated and cited; otherwise the corollary should be weakened to a dense set of s. The abstract does not promise the simultaneous profile result, so the main advertised claims survive. Minor cosmetic issues: a typo in Section 6 (\"for each S\") and a broken sentence in the proof of Example 7.3.\n\nWho this is for. Fractal geometers working on dimensions, point sets, and projections will get value; the methods are reusable. The main results deserve a serious referee, especially Theorem 5.2 and Theorem 6.4. My recommendation: send to peer review, ask the authors to fix or weaken Corollary 7.7. I would cite the main theorems on the sequence property for intermediate dimensions and profiles.","headline":"The sequence property is proved for genuinely new classes (Assouad, intermediate, profiles) with clean compactness/capacity arguments; the only real gap is Corollary 7.7, whose simultaneous profile claim needs continuity or a weaker statement.","tokens_in":14796,"tokens_out":2671,"would_cite":true,"duration_ms":24613,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["28A80","28A78"],"pacs":[],"model":"deepseek-v4-flash","headline":"For many fractal dimensions, a compact set contains a convergent sequence of the same dimension.","keywords":["sequence property","fractal dimension","upper box dimension","Assouad spectrum","intermediate dimensions","dimension profiles","finite stability","dimension points"],"falsifier":"Run the construction of Corollary 7.6 on a self-affine carpet whose intermediate dimension function is known at several parameter values, producing the sequence $(a_i)$, and numerically estimate $\\dim^\\theta\\{a_i\\}$ at those values; the theorem predicts exact equality at every $\\theta$, so any strict shortfall refutes the simultaneous claim. For the single-dimension theorem, take a compact set with known upper box dimension, choose the maximal $r_n$-separated sets used in Theorem 3.2, and check that the enumerated union has exactly that upper box dimension.","tokens_in":13872,"feed_emoji":"📏","tokens_out":12127,"duration_ms":111016,"temperature":0.7,"pith_summary":"For a whole family of fractal dimensions—upper box dimension, Assouad dimension, the Assouad spectrum, upper intermediate dimensions, and box-dimension profiles—the paper proves that every compact set contains a convergent sequence of points whose dimension equals the dimension of the whole set. The proof rests on a localisation lemma: for any dimension that is monotone and finitely stable, some point $x$ of a compact set $E$ has the property that every ball centred at $x$ meets $E$ in a set of full dimension. The same machinery decides when one sequence can witness several dimensions at once: if $E$ is dimension-homogeneous, meaning every ball centred in $E$ contains a bi-Lipschitz copy of $E$, then a single convergent sequence has the same intermediate dimensions, Assouad spectrum, and dimension profiles as $E$ itself. The paper also constructs sets where no single sequence can witness two different dimensions, so the homogeneity condition is not merely technical.","feed_headline":"One sequence can carry a fractal's full dimension","feed_subtitle":"Upper box, Assouad, intermediate and profile dimensions can all be witnessed by one convergent sequence.","key_machinery":"The load-bearing object is the dimension point: a point $x$ in a compact set $E$ such that $\\dim(E \\cap B(x,r)) = \\dim E$ for every $r > 0$. Lemma 2.5 proves such a point exists for every monotone, finitely stable dimension, so the whole construction reduces to placing finite sets of near-full dimension inside the nested balls $B(x, 2^{-n})$ and enumerating their union. A second mechanism, Lemma 7.4, interleaves countably many sequences converging to the same point into one sequence that contains a tail of each, which is what allows one sequence to witness many dimensions at once. The dimension-homogeneous hypothesis—every ball centred in $E$ contains a bi-Lipschitz image of $E$—makes every point of $E$ a common dimension point for all the dimensions considered. Two auxiliary tools carry the individual cases: a capacity estimate showing a finite subset can capture a fixed fraction of a compact set's capacity at a given scale, and the Hausdorff-metric continuity of intermediate pre-measures, which lets dense finite subsets approximate the pre-measure of the whole set.","core_discovery":"On the paper's own terms, the central discovery is that monotone, finitely stable dimensions with the finite-set property satisfy a sequence property: for every non-empty compact $E$ with $\\dim E > 0$ there is a point $x \\in E$ and a sequence $(a_k) \\subset E$ with $a_k \\to x$ and $\\dim(a_k) = \\dim E$. Lemma 2.5 guarantees that every compact set has a dimension point; the proofs then place finite point sets $P_n \\subset B(x, 2^{-n})$ whose $n$-scale dimension is within $1/n$ of the full dimension, enumerate their union as a sequence, and use finite stability to discard finitely many extraneous points. For upper box dimension the finite sets are maximal $r_n$-separated sets; for Assouad dimension and the Assouad spectrum they come from the failure of the defining covering inequality; for intermediate dimensions a Hausdorff-metric continuity lemma for pre-measures produces them; and for dimension profiles a capacity lemma does the same. The final section interleaves the individual witnessing sequences with a tail-combining lemma, so a dimension-homogeneous set has one convergent sequence realising $\\dim^\\theta E$ for all $\\theta \\in (0,1]$ and, in $\\mathbb{R}^n$, realising every profile $\\dim_B^s E$ for $0 < s \\le n$.","pith_inferences":["The authors leave implicit that the same construction scheme should apply to any future monotone, finitely stable dimension with the finite-set property; the proofs for box, Assouad, intermediate and profile dimensions read as templates rather than isolated facts.","A natural test of the limits of the simultaneous result is to ask whether the bi-Lipschitz local structure in the dimension-homogeneity hypothesis can be relaxed to mere existence of a common dimension point; the counterexamples show some condition is needed, but not that full homogeneity is minimal.","Because a single sequence encodes the whole intermediate-dimension function of a homogeneous fractal, one could try to turn the theorem into a numerical recipe: sample the constructed sequence, estimate $\\dim^\\theta$ on the finite point set at several $\\theta$, and use continuity to interpolate across the spectrum.","The sequence property gives a way to think of countable sets as dimension carriers for box-like dimensions, in sharp contrast to Hausdorff dimension, where every countable set has dimension zero; this may clarify how finite stability rather than countability controls the dimension of point sets."],"forward_implications":["Upper box dimension, Assouad dimension, each Assouad spectrum level, each upper intermediate dimension $\\theta \\in (0,1]$, and each box-dimension profile $\\dim_B^s$ satisfy the sequence property on compact sets.","Hausdorff and packing dimensions cannot have the sequence property because they are countably stable, and lower box dimension fails because it is not finitely stable; the paper locates the dividing line at finite stability.","For dimension-homogeneous compact sets, including self-similar, self-affine and graph-directed iterated-function-system attractors, one convergent sequence simultaneously realises every intermediate dimension $\\dim^\\theta E$ and every Assouad spectrum value of $E$.","In $\\mathbb{R}^n$, for such sets a single sequence also realises every profile $\\dim_B^s E$ for $0 < s \\le n$, so the projections of the sequence have almost-surely the same dimensions as the corresponding projections of $E$.","For intermediate dimensions, even without a common dimension point a countable subset of $E$ can share $\\dim^\\theta E$ for all $\\theta \\in (0,1]$ by taking a countable dense set of parameter values and using continuity."],"supporting_citations":[{"why":"established the sequence property for upper box dimension; the paper gives a shorter proof of the same result as a template.","marker":"[18]"},{"why":"defines intermediate dimensions and supplies the continuity in $\\theta$ used to extend simultaneous witnessing from a dense set to all $\\theta$.","marker":"[12]"},{"why":"provides the generalised intermediate-dimension framework and the lower box dimension counterexample showing that finite stability is needed.","marker":"[2]"},{"why":"supplies the standard properties of Assouad dimension and the Assouad spectrum, including finite stability and the spectrum of regular self-similar sets used in the examples.","marker":"[15]"},{"why":"characterises attainable Assouad spectra, used to build the Example 7.1 set where no sequence witnesses both halves of the spectrum.","marker":"[25]"},{"why":"introduces the capacity approach to dimension profiles, which is the definition and finite-stability engine for Section 6.","marker":"[9]"},{"why":"develops the capacity characterisation and projection interpretation of dimension profiles used in Proposition 6.2 and Corollary 7.7.","marker":"[10]"},{"why":"gives the extreme projection-dimension behaviour used in Example 7.3 to separate box dimension from projected dimensions.","marker":"[13]"}],"fun_headline_variants":["Single sequence realizes fractal dimensions","A convergent sequence attains the set's dimension","Sequence property: one sequence, multiple dimensions","Witnessing every fractal dimension with one sequence","One sequence carries many fractal dimensions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The simultaneous witnessing results assume the set is dimension homogeneous—every ball centred in $E$ contains a bi-Lipschitz copy of $E$—so every point of $E$ is a common dimension point for all the dimensions under consideration; without such a common point the paper's examples show that simultaneous witnessing can fail.","fun_headline_variants_meta":{"raw":{"variants":["Single sequence realizes fractal dimensions","A convergent sequence attains the set's dimension","Sequence property: one sequence, multiple dimensions","Witnessing every fractal dimension with one sequence","One sequence carries many fractal dimensions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000641,"raw_usage":{"total_tokens":2944,"prompt_tokens":932,"completion_tokens":2012,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":548,"completion_tokens_details":{"reasoning_tokens":1949}},"tokens_in":548,"tokens_out":2012,"duration_ms":17388,"temperature":1.0,"reasoning_tokens":1949,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T18:35:00.649629+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the construction of Corollary 7.6 on a self-affine carpet whose intermediate dimension function is known at several parameter values, producing the sequence $(a_i)$, and numerically estimate $\\dim^\\theta\\{a_i\\}$ at those values; the theorem predicts exact equality at every $\\theta$, so any strict shortfall refutes the simultaneous claim. For the single-dimension theorem, take a compact set with known upper box dimension, choose the maximal $r_n$-separated sets used in Theorem 3.2, and check that the enumerated union has exactly that upper box dimension.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"established the sequence property for upper box dimension; the paper gives a shorter proof of the same result as a template."},{"cited_title":"Falconer, J.M","cited_arxiv_id":null,"evidence_quote":"defines intermediate dimensions and supplies the continuity in $\\theta$ used to extend simultaneous witnessing from a dense set to all $\\theta$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the generalised intermediate-dimension framework and the lower box dimension counterexample showing that finite stability is needed."},{"cited_title":"Fraser.Assouad Dimension and Fractal Geometry, Cambridge University Press, Cambridge, 2021","cited_arxiv_id":null,"evidence_quote":"supplies the standard properties of Assouad dimension and the Assouad spectrum, including finite stability and the spectrum of regular self-similar sets used in the examples."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"characterises attainable Assouad spectra, used to build the Example 7.1 set where no sequence witnesses both halves of the spectrum."},{"cited_title":"Falconer","cited_arxiv_id":null,"evidence_quote":"introduces the capacity approach to dimension profiles, which is the definition and finite-stability engine for Section 6."},{"cited_title":"Falconer","cited_arxiv_id":null,"evidence_quote":"develops the capacity characterisation and projection interpretation of dimension profiles used in Proposition 6.2 and Corollary 7.7."},{"cited_title":"Falconer and J.D","cited_arxiv_id":null,"evidence_quote":"gives the extreme projection-dimension behaviour used in Example 7.3 to separate box dimension from projected dimensions."}],"review_version":1}