{"id":"71fa408a-d948-457e-a9a1-077685eed392","arxiv_id":"2411.09989","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For a biased aging continuous-time random walk with finite-mean, infinite-variance waiting times, the long-time position distribution is governed by a fractional advection-diffusion equation in space, while its far tail follows an infinite density that depends on the aging time.","lead":"This paper studies a type of random walk with waiting times that have a finite average but an infinite spread, and includes aging effects from the start time. It derives a diffusion equation with a fractional space derivative and shows that rare, slow particles control the far tail of the distribution.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (21) hinges on dropping σ²k² inside the stable factor of Eq. (18); the required smallness condition (σ²/a)(t/tbar)^{-1/α}≪1 is never stated, and the simulations do not probe the regime where it fails.","rationale":"The reader's weakest assumptions is the right one, but it can be sharpened. The paper's main advertised result Eq. (21) rests on dropping σ² from the stable factor while keeping it in the Gaussian factor; although this can be justified at central scaling, the paper does not supply the ordering, and the existing comparisons do not cover the dangerous small-a/large-σ regime. I found no internal inconsistency in the rare-event formulas: Eq. (31) reduces correctly to the ordinary CTRW limit Eq. (34) at ta≪t and to the equilibrium limit Eq. (37) at ta≫t, and Eq. (45) matches the position tail through x∼aN. The explicit admission that Eq. (21) gives an infinite MSD is an honest scope limitation, not a contradiction, because the exact finite-t process has finite moments controlled by rare events outside the central limit. The proposed numerical test would settle whether the two-scale ansatz is actually controlled; until then CONDITIONAL is the appropriate verdict.","tokens_in":17148,"tokens_out":32578,"duration_ms":353588,"concrete_test":"One decisive check is to compare Eq. (23) with the exact Fourier expression obtained from Eq. (18), namely P~(k,ta,t)=∫_0^t ω(ta,t-y) exp{(-ika-σ²k²/2)y/⟨τ⟩} exp{[-ik(a-iσ²k/2)]^α y/tbar}dy, after numerical inverse Fourier transform. Use a stress case far from the tested regime, e.g., α=3/2, τ0=0.1, a=0.1, σ=1, ta=t=10^3 and also 10^6, and compute the normalized L1 difference over the central region ξ∈[-2,2]. If the difference does not decay like t^{-1/α}, or is not within simulation error at t=10^3, the neglected term is not subdominant and Eq. (21) needs an explicit validity condition or amendment.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Eq. (18), the inner ξ-integral is not merely a stable characteristic at k; with the Gaussian variance retained it is exactly exp{[-ik(a - iσ²k/2)]^α y/tbar}. Passing to Eq. (19) replaces this by exp{(-ika)^α y/tbar}, discarding a correction of relative size ~ ασ²k/a in the exponent. At the natural central scale k ~ (t/tbar)^{-1/α}, this correction is O((σ²/a)(t/tbar)^{-1/α}), while the kept Gaussian diffusion term in the first line is O(t^{1-2/α}); for 1<α<2 the kept term is indeed the larger correction, so the step is justifiable under the ordering (σ²/a)(t/tbar)^{-1/α}≪1. The problem is that this ordering is nowhere stated; the phrase \"can be neglected as k→0\" is not a controlled expansion, and the same σ² is retained in the Gaussian factor exp(-σ²k²y/(2⟨τ⟩)). The paper's simulations use a=0.5 or a=1, for which the omitted correction is only percent-level at t=1000, so they do not probe regimes where a is small or σ is large. Equation (21), and its convolution solution (23), are therefore advanced without a stated validity domain for the bias/variance ratio.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the biased aging continuous-time random walk (ACTRW) with power-law waiting times of index 1<α<2 (finite mean, infinite variance) and Gaussian displacements with mean a and variance σ². Using the subordination representation and double Laplace/Fourier transforms, it derives an aging fractional advection-diffusion equation (Eq. (21)) with advection and diffusion constants V=a/⟨τ⟩, D=σ²/(2⟨τ⟩), a spatial fractional derivative with strength S=a^α/t̄, and an aging-dependent source term. It further derives the far tail of the position distribution in the form of an infinite density (Eq. (31)), the corresponding asymptotic MSD (Eq. (40)), and analogous far-tail results for the renewal count (Eqs. (45), (47), (50)). The analytical predictions are compared with simulations for several aging times and for a modified ACTRW with different α in the preparation interval.","tokens_in":17428,"tokens_out":12328,"duration_ms":126089,"significance":"If the asymptotic ordering is made explicit, the paper is a useful contribution. It shows that a fractional space operator can arise from heavy-tailed waiting times rather than heavy-tailed jumps, gives explicit parameter-free transport constants, demonstrates that typical fluctuations are largely aging-insensitive while rare fluctuations are not, and recovers the known limits ta→0 and ta→∞ of [13] and the equilibrium CTRW. The simulation checks use large sample sizes and cover several aging times, and the constants in Eqs. (22), (31), and (45) are expressed in terms of model inputs with no fitted parameters. The main weakness is that the key step from Eq. (18) to Eq. (19) rests on an unstated smallness condition; this is a load-bearing point that needs to be clarified before the central equation can be considered established.","major_comments":[{"comment":"The one-sentence justification for dropping (σ²/2)k² in the inner Lévy-stable factor is not a controlled asymptotic statement, and it is load-bearing for Eq. (21). With the Gaussian variance retained, the inner integral in Eq. (18) is exactly exp{[-ika − σ²k²/2]^α y/t̄} (up to the Fourier convention); passing to Eq. (19) replaces this by exp{(-ika)^α y/t̄}, discarding a correction of relative order α σ² k/a in the exponent. At the natural central scale k ∼ (t/t̄)^{-1/α} this correction is O((σ²/a)(t/t̄)^{-1/α}), which is small only under an ordering such as σ²/a ≪ (t/t̄)^{1/α}. This ordering is nowhere stated, and the same σ² is retained in the Gaussian factor exp(-σ²k²y/(2⟨τ⟩)) in the first line of Eq. (19), so the treatment is asymmetric. The simulations use a=0.5 or 1 with σ=1 at t=1000, for which the omitted correction is only a few percent; they do not probe regimes where a is small or σ is large. Please state the required ordering and justify the two-scale separation, or reformulate Eq. (21) as a leading-order approximation with an explicit error estimate.","section":"Section III, Eq. (18) to Eq. (19)"},{"comment":"The domain of validity of Eq. (45) is not stated precisely enough. The derivation assumes s + ik/⟨τ⟩, u, and s are all small and comparable, which corresponds to ε = N − t/⟨τ⟩ of order t. As written, Eq. (45) is asserted for 'both ta and t are large' without restricting ε. For |ε| ≫ t, the first and third terms in Eq. (45) cancel at leading order, leaving a term proportional to (1−α)/(−⟨τ⟩ε)^α, which is negative for 1<α<2; hence the formula becomes non-positive in the far tail. The paper should explicitly state that Eq. (45) applies in the scaling region ε = O(t), and should not be interpreted as an arbitrarily deep tail.","section":"Section V, Eq. (45)"}],"minor_comments":[{"comment":"In Eq. (23), the stable-density normalization appears to contain an extra factor 1/α after (1/(t/t̄)^{1/α}); this should be removed so that the convolution kernel is the normalized stable density.","section":"Eq. (23)"},{"comment":"The phrase 'Clearly, this prediction is wrong' is too strong: the fractional equation is intended to describe typical fluctuations, and its divergent second moment indicates that moments are not captured by this bulk description rather than that the equation is false. Please rephrase to avoid implying that Eq. (21) is invalidated by this property.","section":"Section III, Eq. (24)"},{"comment":"There are numerous minor typographical errors, such as 'obtained' in the introduction, missing articles in several figure captions, and the caption of Fig. 1 reading 'Comparison Eq. (21) with...' instead of 'Comparison of Eq. (21) with...'. A careful proofread is advised.","section":"General"},{"comment":"The delta-function contribution in Eq. (31) is written with a coefficient that is positive for 1<α<2 (after combining the two negatives), but this is not obvious from the displayed formula; a brief explanation of the origin and sign of the non-moving-particle term would improve readability.","section":"Section IV, Eq. (31)"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the journal's scope and the main results are likely correct, but the missing validity condition at Eq. (18)→(19) is load-bearing for the central equation. The requested revision is local and should be addressable without changing the overall approach; hence major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: this is the first treatment I know of for the biased ACTRW with waiting times of finite mean and infinite variance (1<α<2), and the main results look right. The derivation of the aging fractional advection-diffusion equation (21) is a genuine extension of [13] and [25], not a reparameterization: the aging first-waiting-time kernel enters the Laplace structure non-trivially. The large-deviation tails for position (Eq. 31) and renewal count (Eq. 45) are new for finite ta, and the limiting cases ta→0 and ta→∞ correctly reproduce the earlier CTRW and equilibrium results. The simulations are extensive (up to 10^7 runs) and match the analytic formulas across several aging times. There are no fitted parameters; all constants are expressed through model inputs. That is real evidence.\n\nThe soft spot is the step from (18) to (19), where the Gaussian variance σ²k² is dropped inside the stable Lévy factor while being retained in the ordinary advection-diffusion factor. The paper justifies this with \"as k→0\", which is a heuristic two-scale statement. A controlled expansion shows the omitted correction is of relative order (σ²/a)(t/tbar)^{-1/α}, and since 1<α<2 this vanishes in the long-time limit, so the advertised equation is asymptotically valid. The problem is the ordering is never stated, and the simulations (a=0.5 or 1, σ=1) do not probe regimes where this correction is significant. I would call this a minor presentation gap, not a load-bearing flaw: the central claims about the long-time limiting behavior survive, and the paper honestly acknowledges that Eq. (21) fails for the mean-squared displacement, so the scope is clear.\n\nA few smaller things: there are minor typos, and the derivation of Eq. (27) in Appendix A involves a number of \"ignoring higher-order\" steps that are standard but would benefit from a sentence or two of justification. The delta-function source term in (21) is unusual but well-motivated as the non-moving fraction.\n\nI'd send this to a serious referee. It deserves referee time: the problem is relevant, the methods are standard but applied to a new regime, and the claims are concrete and testable. The referee should ask the authors to state the validity condition for the (18)-(19) step and to add a simulation point with small bias or large σ to demonstrate where the approximation starts to fail. That is a revision, not a rejection.\n\nWho is this for: anyone working on aging renewal processes, fractional kinetic equations, or large deviations in random walks. I'd cite it and bring it to reading group.","headline":"New finite-aging-time results, essentially correct; the main gap is a heuristic two-scale step that should be stated as an asymptotic ordering, not a fatal flaw.","tokens_in":17982,"tokens_out":5074,"would_cite":true,"duration_ms":51236,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60K05","60F10","82C31","26A33"],"pacs":["05.40.Fb","05.60.-k"],"model":"deepseek-v4-flash","headline":"Biased aging random walks with fat-tailed waiting times are shown to obey a fractional advection-diffusion equation in space, with an aging-dependent infinite-density tail describing their rare fluctuations.","keywords":["aging continuous-time random walk","fractional advection-diffusion equation","spatial fractional derivative","infinite density","large deviations","rare fluctuations","renewal process","subordination"],"falsifier":"Numerically invert the unsimplified Fourier expression (18) with the $\\sigma^2 k^2$ term retained in the Levy-stable factor, and compare the result with Eq. (19) for $\\alpha=3/2$, $\\tau_0=0.1$, $t=t_a=1000$, $a=1$, $\\sigma=1$; if the difference does not vanish as $t$ grows, the central fractional equation fails. Independently, the paper's own statement that Eq. (21) yields an infinite second moment can be checked against simulated MSD from the rare-event tail: if the MSD predicted by Eq. (31) deviates from simulations at finite $t_a$, the tail claim is falsified.","tokens_in":16891,"feed_emoji":"⏳","tokens_out":7288,"duration_ms":68837,"temperature":0.7,"pith_summary":"The paper's aim is to show that an aging continuous-time random walk with waiting times that have a finite mean but an infinite variance, and Gaussian jumps with a nonzero bias, is governed in the long-time limit by a fractional advection-diffusion equation whose fractional derivative acts on space, not time. This matters because it means a power-law tail in waiting times alone can produce fractional spatial transport once a bias is present, and because the far tail of the position distribution is an 'infinite density' that remains sensitive to the aging time even when the central part of the distribution has forgotten it. The authors also claim a tight link between the rare-event tails of position and of the number of renewals, and they support the analytic results with simulations. A sympathetic reader would take the central contribution to be the explicit aging fractional equation, its rare-event tail, and the demonstration that these two layers describe both typical and extreme fluctuations.","feed_headline":"Fat waiting times, not jumps, create spatial fractional diffusion","feed_subtitle":"Rare fluctuations of the biased aging walk stay sensitive to aging time and follow an explicit infinite-density tail.","key_machinery":"The load-bearing object is the subordination formula $P(x,t_a,t)=\\sum_N Q_{t_a,t}(N)f(x|N)$, combined with the double-Laplace renewal representation of $Q_{t_a,t}(N)$. In the long-time, small-$k$ limit the key move is to neglect the Gaussian width term $\\sigma^2 k^2$ inside the Levy-stable factor while keeping it in the advection-diffusion factor; this converts the stable factor into $\\exp[(-ika)^\\alpha y/\\bar t]$ and produces the fractional space derivative. The second structural object is the dimensionless variable $\\zeta=1-(x/a)/(t/\\langle\\tau\\rangle)$, which isolates the rare-event region where $x-at/\\langle\\tau\\rangle$ is of order $t$ and turns the tail into the explicit infinite-density forms.","core_discovery":"In the paper's own terms, the central discovery is that the long-time position PDF of the biased ACTRW with $1<\\alpha<2$ solves $$\\frac{\\partial P}{\\partial t}=D\\frac{\\$partial^{2}$ P}{\\partial $x^{2}$}-V\\frac{\\partial P}{\\partial x}+S\\frac{\\partial^\\$\\alpha$ P}{\\partial(-x)^\\$\\alpha$}+\\omega(t_a,t)\\delta(x),$$ with $D=\\sigma^2/(2\\langle\\tau\\rangle)$, $V=a/\\langle\\tau\\rangle$, $S=a^\\alpha/\\bar t$, and with a right Riemann-Liouville fractional derivative in space. The far tail is the infinite density $$P(x,t_a,t)\\sim\\frac{\\tau_0^\\$\\alpha$}{a t^\\$\\alpha$\\$zeta^{{1+\\alpha}}$}\\bigl[\\$\\alpha$+(2-\\$\\alpha$)\\zeta\\bigr]-\\frac{\\tau_0^\\$\\alpha$}{a t^\\$\\alpha$}\\left(\\zeta+\\frac{t_a}{t}\\right)^{-\\$\\alpha$}+\\frac{\\tau_0^\\$\\alpha$}{\\langle\\tau\\rangle(1-\\$\\alpha$)}\\bigl[(t+t_a)^{1-\\$\\alpha$}-$t^{{1-\\alpha}}$\\bigr]\\delta(x),$$ where $\\zeta=1-(x/a)/(t/\\langle\\tau\\rangle)$. The tail is non-integrable, depends on $t_a$ through the ratio $t_a/t$, and controls the mean-squared displacement, whereas the central part of the distribution is aging-insensitive. The same method gives the far tail of the number of renewals, $$Q_{t_a,t}(\\epsilon)\\sim\\tau_0^\\$\\alpha$\\left[\\frac{2-\\$\\alpha$}{(-\\langle\\tau\\rangle\\epsilon)^\\$\\alpha$}+\\frac{\\$\\alpha$ t}{(-\\langle\\tau\\rangle\\epsilon)^{1+\\$\\alpha$}}\\right]-\\frac{\\tau_0^\\$\\alpha$}{(t_a-\\langle\\tau\\rangle\\epsilon)^\\$\\alpha$},$$ tying position and renewal rare events through $x\\sim aN$.","pith_inferences":["A testable consequence the authors leave implicit: for any narrow displacement distribution with finite nonzero mean and variance, not just Gaussian, the same two-scale approximation should yield Eq. (21) with the same $D$, $V$, $S$; a simulation with Laplace or asymmetric jump distributions would check whether the constants remain universal.","The two-scale approximation predicts a crossover: when $\\sigma$ is large or $a$ is small, the neglected $\\sigma^2 k^2$ term in the stable factor may matter at intermediate times; measuring the convergence rate of the central part to the Levy-stable form would quantify when the fractional equation becomes valid.","The delta-source term in Eq. (21) means non-moving particles contribute mass at the origin; a natural extension is the full time-dependent solution for all $t_a$, which would interpolate between the CTRW and equilibrium limits and could be compared with the infinite-density tail at finite $t_a$.","Because the tail is an infinite density, its normalization is not defined; this suggests that any coarse-grained binning in simulations must be handled with care, and that integrated tail probabilities rather than pointwise PDF values are the more robust observable for comparing theory and experiment."],"forward_implications":["The fractional operator in the kinetic equation is spatial even though the jump distribution is Gaussian: a fat-tailed waiting time plus a bias is sufficient, so power-law jumps are not required for fractional space transport.","Typical fluctuations of position follow the same aging-insensitive Levy stable form, while the far tail depends on $t_a/t$; therefore tail measurements can estimate the aging time even when the central part cannot.","The mean displacement is linear, with speed $a/\\langle\\tau\\rangle$, but the second moment is infinite under the fractional equation; the MSD must be computed from the rare-event tail, which gives a $t^{3-\\alpha}$-type growth with $t_a$-dependent corrections.","The far tail of position and the far tail of the number of renewals are governed by the same rare-event mechanism through $x\\sim aN$, so observing one tail predicts the other.","The aging fractional equation also covers the modified ACTRW with different power-law indices before and after the observation start, which is confirmed by simulation."],"supporting_citations":[{"why":"Defines the ACTRW renewal framework and supplies the double-Laplace form for the first waiting time used throughout the derivation.","marker":"[12]"},{"why":"Provides the aging diffusion equation and the renewal representation of Q_{ta,t}(N) in Eq. (8).","marker":"[7]"},{"why":"Gives the fractional advection-diffusion equation for the non-aging CTRW that this work extends to aging.","marker":"[25]"},{"why":"Supplies the infinite-density rare-event theory and the ta=0 and ta->infinity limits that Eqs. (31) and (37) reduce to.","marker":"[13]"},{"why":"Provides the first-waiting-time scaling and the renewal large-deviation tail that Eq. (45) generalizes.","marker":"[19]"},{"why":"Underlies the single-big-jump principle connecting rare position fluctuations to rare renewal counts.","marker":"[33]"},{"why":"Gives the renewal-count distribution Q_s(N) without aging used in the subordination factorisation.","marker":"[37]"},{"why":"Introduces the infinite-density concept used to interpret the non-integrable tail of the positional PDF.","marker":"[60]"}],"fun_headline_variants":["Rare events set the infinite-density tail of aging walks","Space-fractional diffusion from fat waiting times alone","Aging walk's far tail answers to renewals, not typical spread","Biased aging CTRW: rare fluctuations govern long-time spread"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The derivation assumes that, at long times, the finite width of the jump lengths stops mattering inside the fractional part of the equation even though it still sets the ordinary diffusion constant; this two-scale split is the load-bearing approximation and is not derived from a controlled expansion.","fun_headline_variants_meta":{"raw":{"variants":["Rare events set the infinite-density tail of aging walks","Space-fractional diffusion from fat waiting times alone","Aging walk's far tail answers to renewals, not typical spread","Biased aging CTRW: rare fluctuations govern long-time spread"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000332,"raw_usage":{"total_tokens":1883,"prompt_tokens":1021,"completion_tokens":862,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":637,"completion_tokens_details":{"reasoning_tokens":792}},"tokens_in":637,"tokens_out":862,"duration_ms":8684,"temperature":1.0,"reasoning_tokens":792,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T20:05:05.139488+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Numerically invert the unsimplified Fourier expression (18) with the $\\sigma^2 k^2$ term retained in the Levy-stable factor, and compare the result with Eq. (19) for $\\alpha=3/2$, $\\tau_0=0.1$, $t=t_a=1000$, $a=1$, $\\sigma=1$; if the difference does not vanish as $t$ grows, the central fractional equation fails. Independently, the paper's own statement that Eq. (21) yields an infinite second moment can be checked against simulated MSD from the rare-event tail: if the MSD predicted by Eq. (31) deviates from simulations at finite $t_a$, the tail claim is falsified.","supporting_citations":[{"cited_title":"The inverse Fourier-Laplace transform is straightforward for this term","cited_arxiv_id":null,"evidence_quote":"Gives the fractional advection-diffusion equation for the non-aging CTRW that this work extends to aging."},{"cited_title":"The key idea, based on Eq","cited_arxiv_id":null,"evidence_quote":"Supplies the infinite-density rare-event theory and the ta=0 and ta->infinity limits that Eqs. (31) and (37) reduce to."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the first-waiting-time scaling and the renewal large-deviation tail that Eq. (45) generalizes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Underlies the single-big-jump principle connecting rare position fluctuations to rare renewal counts."},{"cited_title":"Taking ta → ∞ , Eq","cited_arxiv_id":null,"evidence_quote":"Gives the renewal-count distribution Q_s(N) without aging used in the subordination factorisation."}],"review_version":1}