{"id":"305362bd-4ab8-41bc-930a-7da9d3eef34e","arxiv_id":"2411.19186","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Interplay of inertial and active time scales in a run-and-tumble particle yields four dynamical regimes, analytic position distributions, and a long-time large deviation function.","lead":"An inertial run-and-tumble particle, with both a velocity relaxation time and an active tumbling time, moves on a line through four distinct dynamical regimes with different growth of its spread. The paper derives the position distributions and a long-time large deviation function, relevant for macro-scale active particles like vibrobots or hexbugs.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The inertia-dominated survival scaling Eq. (88) is inferred from a narrow data collapse and appears inconsistent with the known random-acceleration scaling of regime R3, so the claimed m^{1/12} amplitude is not established.","rationale":"The paper's main contributions—the exact moment recursion, the four-regime classification, the MSD asymptotics, and the large-deviation function Φ(w)=1−√(1−w^2)—are supported by analytic derivations and consistent numerical data. The MSD results in Table I follow from the exact recursion relations, and the LDF in Sec. V D is obtained through a coarse-graining argument whose saddle point is internally consistent. These parts of the central claim appear sound. The load-bearing weakness is confined to the first-passage section, where Eq. (88) postulates a scaling form with an exponent 1/12 that is extracted from data collapse rather than derived. The reader's weakest_assumption correctly identifies this point. My stress-test strengthens it: a scale-invariance argument for the random-acceleration limit in regime R3 predicts a different amplitude exponent (m^{1/6} and x0^{1/6}), so Eq. (88) is not just underived but potentially incorrect. This matters because the abstract claims the persistence exponents are found theoretically; the exponents t^{-1/4} and t^{-1/2} are themselves supported by known results, but the amplitude scaling is presented as a result without proof. The proposed numerical test—varying m and x0 over a wide range and measuring the amplitude exponent—would settle whether Eq. (88) holds or should be revised. Since the central regime and distribution results are unaffected, the appropriate verdict remains conditional rather than reject or accept. The reader's CONDITIONAL verdict is therefore appropriate, and no change to it is needed from this stress-test pass.","tokens_in":19060,"tokens_out":26834,"duration_ms":216750,"concrete_test":"Simulate the IRTP in the deep R3 regime: fix x0 = 1, a0 = 1, τa = 0.2, and choose an observation time t = 10 τa with τm ≫ t (e.g., γ = 0.01 and m ranging from 1 to 100, so τm = 100·m ranges from 100 to 10^4, keeping t/τm ≤ 0.02). For each m, measure the survival probability Q(t) with high statistics (≥10^6 trajectories). Plot ln Q(t) versus ln m and fit the slope. Random-acceleration scaling predicts slope 1/6; Eq. (90) predicts slope 1/12. Independently, repeat with several x0 values (0.5, 1, 2, 5) at fixed m and extract the x0-exponent: expected 1/6 from random-acceleration scaling versus 1/12 from Eq. (88). A slope close to 1/6 would falsify the proposed Eq. (88) and require correction of the persistence amplitude claims.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim includes persistence exponents found 'theoretically,' but the key amplitude scaling in the inertia-dominated regime is not derived: Eq. (88) postulates Q(t)=C(x0 τa γ/(a0 τm^2))^{1/12} F_m(t/τm), with the 1/12 exponent read off from the data collapse in Fig. 12. This is more than a gap in derivation: in regime R3 (τa ≪ t ≪ τm) the dynamics reduces to m x¨ ≈ a0 σ(t), and for t ≫ τa the dichotomous acceleration is effectively white noise with D = a0^2 τa/(2m^2). For a random-acceleration process x¨ = ξ with ⟨ξξ⟩ = 2Dδ(t), scale invariance forces Q(t) = F(x0/(√D t^{3/2})). Since the persistence exponent is 1/4, F(z) ~ z^{1/6} for small z, giving Q ~ x0^{1/6} D^{-1/12} t^{-1/4} ∝ m^{1/6} t^{-1/4}. Equation (90), however, claims Q ∼ m^{1/12} t^{-1/4}. The m-exponents differ by a factor of two, and the x0-dependences (1/6 vs 1/12) also differ. Thus Eq. (88) is not merely an unproved ansatz; it is in tension with an established scaling law. The numerical collapse over m = 10, 15, 20 (Fig. 12b) spans only a factor of two in m, where m^{1/12} and m^{1/6} differ by about 6% versus 12%, likely too close to distinguish reliably. This is the load-bearing weak point for the persistence claims.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper analyzes a one-dimensional inertial run-and-tumble particle described by m \\dot v = -\\gamma v + a0 \\sigma(t). Starting from the Fokker-Planck equations, the authors derive exact recursions for the moments M(k,n)=<x^k v^n> and obtain closed-form expressions for <v^2>, <xv>, and <x^2>. From these they identify four dynamical regimes R1-R4 with MSD growth t^4, t^2, t^3, and t, respectively, and a late-time effective diffusion constant D_eff = a0^2 \\tau_a/(2\\gamma^2). The paper then computes approximate position distributions in R1 (trajectory expansion in t/\\tau_a), R2 (effective overdamped RTP with a small-mass shift), R3 (dichotomous-acceleration large deviation function), and R4 (large deviation function Phi(w)=1-sqrt(1-w^2) with finite-time corrections). Finally it studies survival probabilities, predicting t^{-1/2} decay in the activity-dominated case and a t^{-1/4} to t^{-1/2} crossover in the inertia-dominated case, with the empirical scaling form (88). All results are compared with numerical simulations.","tokens_in":19504,"tokens_out":25783,"duration_ms":214572,"significance":"The exact moment recursion in Eqs. (9)-(10) is a clean and useful result, and the explicit MSD formulas in Eqs. (20)-(22) reproduce the four regimes with no adjustable parameters. The R4 large deviation calculation is a nontrivial analytic derivation that matches the simulations, and the R1/R2/R3 distribution computations correctly reduce to known overdamped or force-free limits. The paper is honest about using simulations for the inertia-dominated survival scaling. However, the first-passage amplitude claim in Eqs. (88)-(91) is not a consequence of the equations of motion and is inconsistent with the random-acceleration scaling of regime R3, so the persistence section requires substantive revision before the paper can be accepted.","major_comments":[{"comment":"The scaling form (88) is not consistent with the dynamics in regime R3. For \\tau_a << t << \\tau_m, Eq. (58) reduces to m x¨ ≈ a0 \\sigma(t), and for t >> \\tau_a the dichotomous acceleration is effectively white noise with <\\xi(t)\\xi(t')> = 2D\\delta(t-t'), where D = a0^2 \\tau_a/(2m^2). Scale invariance of this random-acceleration process forces Q(t) = F(x0/(\\sqrt{D} t^{3/2})); with the known persistence exponent 1/4, this gives Q(t) ~ C x0^{1/6} D^{-1/12} t^{-1/4} ∝ m^{1/6} t^{-1/4} at fixed \\gamma. This contradicts Eq. (90), which predicts Q ~ m^{1/12} t^{-1/4}; the m and x0 exponents both differ by a factor of two. In addition, Eq. (90) is unphysical in the limit m → ∞ at fixed t: A = x0\\tau_a\\gamma/(a0\\tau_m^2) vanishes but F_m ~ (t/\\tau_m)^{-1/4} ∝ m^{1/4}, so the right-hand side grows as m^{1/12} and would exceed unity, whereas the exact displacement vanishes and Q → 1. The data collapse in Fig. 12(b) spans only m = 10, 15, 20, a factor of two in m, where m^{1/12} and m^{1/6} are too close to distinguish. Since the 1/12 exponent is read off from that collapse rather than derived, the claims in Eqs. (88)-(91) are not established; the authors should derive the correct scaling from Eq. (58) or restrict the claims to the persistence exponents.","section":"Sec. VI, Eqs. (88)-(91)"},{"comment":"The coarse-graining derivation of the long-time large deviation function is performed under the explicit assumption \\alpha = \\tau_a/\\tau_m << 1, i.e., \\tau_a << \\tau_m, but the claimed result (62) is stated for the full regime R4, t >> max(\\tau_m, \\tau_a). For the complementary ordering \\tau_m << \\tau_a the coarse-graining argument does not apply. The final LDF is the same as that of the overdamped RTP, so the result is plausible, but the paper should either supply the argument for the activity-dominated branch or explicitly restrict the derivation; otherwise the theoretical scope of Eq. (84) is narrower than the text claims.","section":"Sec. V D, Eqs. (63)-(84)"}],"minor_comments":[{"comment":"The sentence says the d=2 diagonal equations are solved starting from M(2,0,t), but the coupled structure requires solving M(0,2,t) first; please correct the displayed equation or the solution order.","section":"Sec. III, after Eq. (19)"},{"comment":"Please specify the initial velocity and initial orientation \\sigma when the particle starts at x0; the scaling forms (86) and (88) and the simulations in Fig. 11 are otherwise under-specified.","section":"Sec. VI"},{"comment":"The y-axis labels in Fig. 12 are ambiguous in the current rendering; please ensure that the exponents shown (e.g., m^{1/6} versus m^{1/12} or m^{-1/12}) match the claimed collapse variables.","section":"Fig. 12"},{"comment":"The header contains a typo: 'Run-and-T umble Particle' should read 'Run-and-Tumble Particle'.","section":"Title page"},{"comment":"The dimensionless combination A = x0\\tau_a\\gamma/(a0\\tau_m^2) is introduced without explanation; a short scaling argument or a comment on its physical origin would help the reader understand why this particular combination is chosen.","section":"Eq. (88)"}],"recommendation":"major_revision","confidential_remarks":"This is a strong paper outside Section VI: the moment recursions, the four-regime MSD picture, and the R4 large deviation calculation are sound and valuable. The main risk is that the empirical first-passage scaling (88) will be quoted as an exact result despite being inconsistent with the random-acceleration limit. I recommend major revision rather than rejection, because the correct scaling can be derived and the rest of the paper stands. Please also ask the authors to state the initial conditions used for the survival-probability simulations."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper delivers real value on the position distribution side. The moment recursion (Eqs. 9-10) is clean and exact; the four-regime MSD classification follows correctly, and the R1 trajectory expansion, the R2 small-mass correction, and the R4 large-deviation derivation are all controlled and match simulations well. The R4 LDF reducing to the overdamped RTP result is the right limit, and the derivation is self-contained rather than assumed. For those working on inertial active particles, this is a useful and mostly reliable piece of theory.\n\nThe soft spot is the survival probability in the inertia-dominated case, specifically Eq. (88) and the claimed m^{1/12} amplitude. The 1/12 exponent is read off from a data collapse, not derived, and the abstract's \"theoretically\" overstates this. More worrying, the scaling is in tension with the known random-acceleration limit. In R3 (τa << t << τm), m x¨ ≈ a0 σ(t), and for t >> τa the dichotomous noise becomes effectively white with D = a0^2 τa/(2m^2). For a random-acceleration process, scale invariance forces Q(t) = F(x0/√(D t^3)) with F(z) ~ z^{1/6}, giving Q ~ m^{1/6} t^{-1/4} and x0^{1/6}, not m^{1/12} and x0^{1/12}. That is a factor of two in the m and x0 exponents. The data collapse in Fig. 12(b) spans m = 10 to 20, only a factor of two, where m^{1/12} and m^{1/6} differ by about 6% and 12% respectively — too close to distinguish reliably. So the persistence-exponent claim is on shakier ground than the abstract suggests.\n\nThis does not sink the rest of the paper. The position distributions, MSD, and large-deviation function are solid and worth having. But the survival section needs either a derivation of the correct prefactor scaling from the random-acceleration mapping or a clear acknowledgment that the amplitude exponents are empirical and under-resolved. A referee should ask for that revision.\n\nI would send this to peer review. The central results are correct and useful, and the survival issue is local and fixable. The authors know the R3 mapping; they just need to follow it through for the prefactor.","headline":"Solid analytic work on the inertial RTP position distributions, but the inertia-dominated survival scaling is an unproven fitted exponent that likely conflicts with the random-acceleration limit.","tokens_in":19990,"tokens_out":7669,"would_cite":true,"duration_ms":75783,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Adding inertia to a run-and-tumble particle creates four distinct motion regimes.","keywords":["inertial run-and-tumble particle","active matter","underdamped Langevin equation","mean-squared displacement","position distribution","large deviation function","persistence exponent","survival probability"],"falsifier":"Simulate the inertia-dominated case with $m$, $\\gamma$, and $x_0$ varied independently over at least two decades and check whether $Q(t)m^{-1/12}t^{1/4}$ is flat in the intermediate window and whether all curves collapse as $F_m(t/\\tau_m)$; a systematic drift or a different best-fit exponent falsifies Eq. (88), while measuring $P(x,t)$ for several masses should confirm the $m$-independent LDF $\\Phi(w)=1-\\sqrt{1-w^2}$ at long times.","tokens_in":18847,"feed_emoji":"📈","tokens_out":8807,"duration_ms":72741,"temperature":0.7,"pith_summary":"This paper studies an inertial run-and-tumble particle (IRTP) on a line, described by an underdamped Langevin equation with mass $m$, damping $\\gamma$, and a dichotomous active force that flips at rate $\\tau_a^{-1}$. The authors aim to show that the two intrinsic time scales, inertial $\\tau_m=m/\\gamma$ and active $\\tau_a$, organize the dynamics into four regimes in which the mean-squared displacement grows as $t^4$, $t^2$, $t^3$, and $t$, respectively. They further claim that the position distribution can be computed analytically in each separated-timescale regime, and that at long times it takes a large deviation form $P(x,t)\\sim \\exp[-(t/\\tau_a)\\Phi(\\gamma x/(a_0 t))]$ with $\\Phi(w)=1-\\sqrt{1-w^2}$. The analysis also predicts persistence exponents $t^{-1/4}$ and $t^{-1/2}$ for survival probabilities in the intermediate and long-time regimes. This matters because larger active agents, unlike micron-scale swimmers, retain inertia, and the paper provides exact statistical predictions for their motion that are testable in experiments and simulations.","feed_headline":"Inertia creates four regimes, from t^4 to t^2 to t^3 to t","feed_subtitle":"Mean-squared displacement and position distribution are computed exactly in every regime.","key_machinery":"The argument is carried by three tools. First, the Fokker-Planck equations for the two noise states are recast into a recursive hierarchy for moments $M(k,n,t)=\\langle x^k v^n\\rangle$, whose triangular structure (each diagonal solved sequentially) yields exact lower-order correlations including the MSD. Second, a trajectory-based expansion in powers of $1/\\tau_a$, counting zero, one, or more tumbling events, gives the short-time position distribution; effective mappings reduce the intermediate regimes to known exactly solvable processes (an overdamped RTP for $\\tau_m\\ll t\\ll\\tau_a$, and a dichotomous acceleration process for $\\tau_a\\ll t\\ll\\tau_m$). Third, the long-time regime is handled by coarse-graining the dichotomous noise over time windows much longer than $\\tau_a$, which yields an effective noise with large deviation function $S(w)=1-\\sqrt{1-w^2}$, and the position LDF follows from a saddle-point evaluation and Legendre transform. The survival-probability results rely on known persistence exponents of the random acceleration process together with a scale-invariance ansatz.","core_discovery":"On its own terms, the central discovery is that adding inertia to a run-and-tumble particle does not merely smooth the overdamped picture; it creates a qualitatively richer set of scaling laws. The MSD is computed exactly from a recursive moment hierarchy and crosses $t^4$ (short-time ballistic), then either $t^2$ (activity-dominated intermediate regime $\\tau_m\\ll t\\ll\\tau_a$) or $t^3$ (inertia-dominated intermediate regime $\\tau_a\\ll t\\ll\\tau_m$), before reaching normal diffusion $2D_{\\mathrm{eff}}t$ with $D_{\\mathrm{eff}}=a_0^2\\tau_a/(2\\gamma^2)$. The corresponding position distributions are obtained analytically by mapping regime R2 to an overdamped RTP and regime R3 to a dichotomous acceleration process; at long times the large deviation function is $\\Phi(w)=1-\\sqrt{1-w^2}$ for $w=\\gamma x/(a_0 t)$, the same rate function as an overdamped RTP. The survival probability in the inertia-dominated case is claimed to obey $Q(t)=C[x_0\\tau_a\\gamma/(a_0\\tau_m^2)]^{1/12}F_m(t/\\tau_m)$, with a crossover from $t^{-1/4}$ to $t^{-1/2}$.","pith_inferences":["If the scaling ansatz of Eq. (88) holds beyond the simulated range, the crossover time from $t^{-1/4}$ to $t^{-1/2}$ in the inertia-dominated regime should shift with mass and damping in a way one can read off by matching the two power laws; a dedicated simulation scan over $\\tau_a/\\tau_m$ would make this quantitative.","The equality of the late-time LDF with that of an overdamped RTP suggests the rate function is universal across inertia strengths once $x$ is scaled by $a_0t/\\gamma$; one could test whether finite-$m$ corrections collapse onto $\\Phi$ with a single correction exponent.","A similar four-regime structure should appear in underdamped active Brownian particles, where the internal direction diffuses instead of tumbling; the same moment hierarchy could be adapted, and comparing the two would show which features are universal to inertial active motion."],"forward_implications":["The exact MSD formulas imply a crossover sequence $t^4\\to t^2\\to t^3\\to t$ when $\\tau_m\\ll\\tau_a$, and $t^4\\to t^3\\to t^2\\to t$ when $\\tau_a\\ll\\tau_m$; which intermediate law is observed tells which time scale dominates.","In the inertia-dominated intermediate regime the position distribution is confined to the light cone $x\\in[-a_0t^2/m,a_0t^2/m]$ and follows the large deviation form of a dichotomous acceleration process, so measuring $P(x,t)$ there directly tests the effective mapping.","At late times, typical fluctuations are Gaussian with $D_{\\mathrm{eff}}$, while atypical fluctuations obey $\\Phi(w)=1-\\sqrt{1-w^2}$; the kurtosis decays as $1/t$, a signature visible in experiments.","The survival probability in the inertia-dominated regime is predicted to cross from $t^{-1/4}$ to $t^{-1/2}$ with an amplitude depending on $m^{1/12}$ and $\\gamma^{-1/4}$; this is a sharp, testable prediction."],"supporting_citations":[{"why":"Supplies the exact position distribution of the overdamped RTP, the base model used for regime R2 and for the long-time LDF comparison.","marker":"[12,13]"},{"why":"Provides the survival-probability crossover form $F_a(u)$ with $u^{-1/2}$ tail used in the activity-dominated case.","marker":"[11]"},{"why":"Supplies the coarse-grained trajectory formalism and saddle-point method used to derive the long-time LDF.","marker":"[42]"},{"why":"Provides the dichotomous-acceleration position LDF and light-cone support used to describe regime R3.","marker":"[36,41]"},{"why":"Gives the $t^{-1/4}$ persistence exponent of the random acceleration process used to predict the inertia-dominated survival decay.","marker":"[43,44]"},{"why":"Gives the $t^{-1/2}$ persistence exponent of diffusive motion used for the late-time survival decay.","marker":"[45,46]"},{"why":"Provides the stationary velocity distribution of an overdamped RTP in a harmonic trap, to which the IRTP velocity process maps.","marker":"[37]"},{"why":"Provides the dichotomous-noise autocorrelation used in the direct computation of the second-order correlations.","marker":"[47]"}],"fun_headline_variants":["Inertial run-and-tumble: four regimes, exact MSD scalings","From t^4 to t: inertia rewrites run-and-tumble scaling","Inertia yields t^3 regime and exact large deviation function","Four regimes from inertia: t^4, t^2, t^3, and t","Inertia adds a t^3 regime to run-and-tumble dynamics"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The inertia-dominated survival probability is assumed to obey the scaling form $Q(t)=C[x_0\\tau_a\\gamma/(a_0\\tau_m^2)]^{1/12}F_m(t/\\tau_m)$ with the exponent $1/12$ taken from data collapse rather than derived from the equations; if that form does not hold beyond the simulated parameters, the claimed crossover and amplitude dependence are not established.","fun_headline_variants_meta":{"raw":{"variants":["Inertial run-and-tumble: four regimes, exact MSD scalings","From t^4 to t: inertia rewrites run-and-tumble scaling","Inertia yields t^3 regime and exact large deviation function","Four regimes from inertia: t^4, t^2, t^3, and t","Inertia adds a t^3 regime to run-and-tumble dynamics"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000859,"raw_usage":{"total_tokens":3719,"prompt_tokens":928,"completion_tokens":2791,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":544,"completion_tokens_details":{"reasoning_tokens":2687}},"tokens_in":544,"tokens_out":2791,"duration_ms":18303,"temperature":1.0,"reasoning_tokens":2687,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:26:30.023835+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the inertia-dominated case with $m$, $\\gamma$, and $x_0$ varied independently over at least two decades and check whether $Q(t)m^{-1/12}t^{1/4}$ is flat in the intermediate window and whether all curves collapse as $F_m(t/\\tau_m)$; a systematic drift or a different best-fit exponent falsifies Eq. (88), while measuring $P(x,t)$ for several masses should confirm the $m$-independent LDF $\\Phi(w)=1-\\sqrt{1-w^2}$ at long times.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the coarse-grained trajectory formalism and saddle-point method used to derive the long-time LDF."},{"cited_title":"Adersh , author M","cited_arxiv_id":null,"evidence_quote":"Provides the stationary velocity distribution of an overdamped RTP in a harmonic trap, to which the IRTP velocity process maps."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the dichotomous-noise autocorrelation used in the direct computation of the second-order correlations."}],"review_version":1}