Beyond Poisson: First-Passage Asymptotics of Renewal Shot Noise
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.
Kai: We're looking at the paper: "Beyond Poisson: First-Passage Asymptotics of Renewal Shot Noise".
Title and authors: Kai: The paper, 'Beyond Poisson: First-Passage Asymptotics of Renewal Shot Noise', is written by J Brémon, and it comes from a college de france lab. It’s tackling the limitation where analytical solutions were only available for the Poisson case.
Mira: It's really about extending that understanding to situations where arrivals aren't random in the simple Markovian way; they can follow other renewal statistics, not just Poisson.
Lev: I wonder how this translates to our error correction work; if we have correlated noise, can this formula give us a better estimate for when an error-free state is reached?
Kai: The paper focuses on providing a universal asymptotic formula for the mean first-passage time to reach a level b, no matter what the arrival statistics are.
Mira: That universality is key; it suggests that the physics governing these crossings is dictated by simpler parameters, specifically how fast the arrivals decay and their short-time behavior.
Lev: If it’s universal, it means we don't need a specific model for every single system we build; there's one mathematical structure that describes the scaling.
The paper's summary: Kai: The summary says they derived this simple, closed-form expression for the MFPT, which is what really opens things up because it’s exact asymptotically.
Mira: They achieved this by finding a new way to calculate the Laplace transform of all the moments of X t, which was previously unavailable in the literature. That's a significant mathematical step forward.
Lev: So it’s not just an approximation anymore; they have an exact expression for these moments, which is huge for testing any model we try to use on real hardware.
Kai: The physical interpretation they give is that temporal correlations modulate the baseline Arrhenius scaling directly; the bursts of arrivals are what cause the deviation from that simple exponential trend.
Mira: They show this explicitly by linking it to the shape parameter k and the mean rate r of a Gamma distribution describing those interarrival times, as shown in Figure two <ref:2510.21670#pg2>.
Lev: When they look at the scaling, they mention that if short gaps are prevalent, meaning k is less than or equal to one, crossings are driven by bursts rather than isolated events.
The paper's improvements: Kai: One of the main improvements they point out is that for large thresholds b, the full first-passage time distribution becomes exponential at large enough values of b.
Mira: That confirms that the MFPT value you get from this formula is a complete asymptotic characterization because P t greater than t approaches exp negative t over the mean MFPT as b goes to infinity.
Lev: That convergence to the rare-event estimate, which they call TbRE, is important because it tells us when we can trust that theoretical prediction for extremely rare events in a system we might actually be simulating.
Kai: They also show how this relationship between the actual MFPT and that rare-event estimate changes depending on the burstiness parameter kappa.
Mira: Specifically, they confirm that the convergence of the ratio Tb over TbRE to one is slower in regimes where kappa is less than or equal to one because of enhanced clustering.
Lev: That means when we are dealing with those clustered events, we need to be more careful about how quickly our theoretical estimates match what we actually measure in a simulation.
Conclusion: Kai: So, summarizing the 'Beyond Poisson: First-Passage Asymptotics of Renewal Shot Noise', they’ve given us a universal formula that handles arbitrary arrival statistics and exponential marks.
Mira: The central result is that temporal correlations universally modulate the baseline Arrhenius law, showing how bursty arrivals accelerate threshold crossings.
Lev: For me, the main practical implication is getting an exact way to calculate these moments so we can run better simulations for real hardware overheads in systems with correlated noise.
Kai: They also confirmed that the distribution becomes exponential at large thresholds, which gives us a solid asymptotic characterization of the MFPT value.
Mira: And they explicitly showed that finite threshold corrections beyond the rare-event estimate are quantifiable when we look at regimes where temporal correlations cause enhanced clustering.
Lev: I think for us, this means we have a more robust mathematical tool to analyze how bursty dynamics affect the performance limits of error correction codes in noisy environments.
J. Brémont
Collge de France
cond-mat.stat-mech, math.PR
Submitted: 2025-10-24
Updated: 2026-02-23
Comments: PRL version
Journal ref: Phys. Rev. Lett. 136, 107101 (2026)
DOI: 10.1103/bbh8-n8dt
Code: https://github.com/julien-bremont/Shot-noise
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 90/100
The gist: The paper derives a universal asymptotic formula for the mean first-passage time (MFPT) of renewal shot noise with arbitrary arrival statistics and exponential marks…
Key concepts
- Mean First-Passage Time (MFPT) ⟨Tb⟩
- This is the average time it takes for a renewal shot noise process to reach or exceed a specific threshold 'b'. The paper derives an exact asymptotic formula for this value, which describes the long-term behavior of how quickly the system crosses that barrier.
- Arrhenius Law Scaling
- The baseline scaling for MFPT in many systems is governed by an Arrhenius law, meaning the time required grows exponentially with the threshold. The paper shows that temporal correlations in arrivals modulate this baseline exponential growth, quantifying how these correlations change the crossing time.
- Temporal Correlations (κ)
- This parameter describes how correlated the arrival times of shots are. When short gaps between arrivals are prevalent (κ ≤ 1), crossings are driven by bursts of arrivals rather than isolated events, leading to a significant acceleration in threshold crossings.
Terminology
Summary
The paper derives a universal asymptotic formula for the mean first-passage time (MFPT) of renewal shot noise with arbitrary arrival statistics and exponential marks, overcoming previous limitations that were confined to the Poisson case The central result is a simple, closed-form expression that exposes how temporal correlations in arrivals modulate the baseline Arrhenius law
Universal Asymptotic Formula
The authors present a simple, exact asymptotic expression for the MFPT ⟨Tb⟩ valid for general renewal arrivals and exponential marks This key formula is given by a compact product formula Eq. (1) This result is enabled by a novel closed-form expression for the Laplace transform of all moments of X(t)
Physical Interpretation of Scaling
The product form in Eq. (1) reflects a clear physical mechanism, where temporal correlations universally modulate the baseline Arrhenius scaling For large thresholds b, the MFPT retains an Arrhenius-like scale ⟨Tb⟩ ∝ eλb, but the product term quantifies how temporal correlations impact this baseline exponential scale The scaling itself is universal and dictated solely by the short-time behavior of w(t) through parameters c and κ When short gaps are prevalent (κ ≤ 1), crossings are driven by bursts of arrivals rather than isolated events, leading to substantial acceleration of threshold crossings
Asymptotic Behavior and Distribution
The full FPT distribution becomes exponential at large thresholds, implying that ⟨Tb⟩ provides a complete asymptotic characterization Specifically, P(Tb > t) ∼ exp(-t/⟨Tb⟩) as b → ∞ This exponential form is confirmed numerically in FIG. 3 The convergence of the actual MFPT ⟨Tb⟩ to the rare-event estimate ⟨Tb⟩RE = 1/(rp(b)) holds for large b
Exact Moment Expressions
The authors provide exact closed-form expressions for all moments of renewal shot noise, which are previously unavailable in the literature For exponentially distributed marks, the stationary moments are given by ⟨Xn∞⟩ → t→∞ ⟨Xn∞⟩ = rλ−n (n − 1)! γ nY−1 m=1 1 + ψˆ(mγ)
Numerical Validation and Scope
The numerical simulations confirm the theoretical results, showing excellent agreement between the theoretical MFPT predicted by Eq. (1) and simulation data The analysis is valid for renewal laws with exponential or faster decay, including Poisson and Gamma renewals The study also provides a link between shot noise and the G/M/∞ queue model
Code Availability
The complete simulation code used to generate the data in FIGs. 2 and 3, along with a Mathematica notebook demonstrating how to evaluate numerically the MFPT ⟨Tb⟩ for nonintegers values of λb in the case of Gamma interarrivals, is publicly available The code can be found at https://github. p. 4com/julien-bremont/Shot-noise
References
[1] W. Gerstner and W. p. 4 M. Kistler, Spiking Neuron Models: Single Neurons, Populations, Plasticity (Cambridge University Press, 2002) [2] A. p. 4 N. Burkitt, Biological Cybernetics 95, 1 (2006) [3] F. p. 4 Droste and B. p. 4 Lindner, Journal of Computational Neuroscience 43, 81 (2017) [4] K. p. 4 Rijal, N. I. C. M¨uller, E. p. 4 Friauf, A. Singh, A. Prasad, and D. p. 4 Das, Physical Review Letters 132, 228401 (2024) [5] P. p. 4 Robin, N. Kavokine, and L. p. 4 Bocquet, Science 373, 687 (2021) [6] J. p. 4 Peccoud and B. p. 4 Ycart, Theoretical Population Biology 48, 222 (1995) [7] N. p. 4 Friedman, L. p. 4 Cai, and X. p. 4 S. Xie, Phys. Rev. Lett. 97, 168302 (2006) [8] V. p. 4 Shahrezaei and P. p. 4 S Swain, Proc Natl Acad Sci U S A 105, 17256 (2008) [9] I. p. 4 Golding, J. p. 4 Paulsson, S. p. 4 M Zawilski, and E. p. 4 C Cox, Cell 123, 1025 (2005) [10] A. p. 4 Raj and A van Oudenaarden, Cell 135, 216 (2008) [11] B Meyer, O B´enichou, Y Kafri, and R Voituriez, Biophysical Journal 102, 2186 (2012/5/2/) [12] J Szavits-Nossan and R Grima, Biophysical Journal 123, 1034 (2024) [13] T Zhu, J Li, A Samanta, A Leach, and K Gall, Physical Review Letters 100, 025502 (2008) [14] P Sollich, F Lequeux, P Hebraud, and M E Cates, Physical Review Letters 78, 2020 (1997) [15] S O Rice, Bell System Technical Journal 23, 282 (1944) [16] G L´eveill´e and E Hamel, Scandinavian Actuarial Journal, 228 (1919) [17] K Borovkov and A Novikov, Statistics & Probability Letters 78, 1517 (2008) [18] O Kella and W Stadje, Journal of Applied Probability 38, 781 (2001) [19] F Laio, A Porporato, L Ridolfi, and I RodriguezIturbe, Physical Review E 63, 036105 (2001) [20] A Tsurui and S Osaki, Stochastic Processes and their Applications 4, 79 (1976) [21] J Masoliver, Physical Review A 35, 3918 (1987) [22] M Deger, M Helias, C Boucsein, and S Rotter, Journal of Computational Neuroscience 32, 443 (2012) [23] N Kumar, A Singh, and R V Kulkarni, PLOS Computational Biology 11, e1004292 (2015) [24] A Nicolas, E E Ferrero, K Martens, and J L Barrat, Reviews of Modern Physics 90, 045006 (2018) [25] K S Olsen and H Löwen, Journal of Physics A: Mathematical and Theoretical 57, 485001 (2024) [3] J Szavits-Nossan and R Grima, Biophysical Journal 123, 1034 (2024) [4] L Tak´acs, Acta Mathematica Academiae Scientiarum Hungarica 9, 45 (1958) [5] M Abramowitz and I Stegun, Handbook of Mathematical Functions: With Formulas, Graphs, and Mathematical Tables, Applied Mathematics Series (Dover Publications, 1965) [6] J Jang, A Dassios, and H Zhao, Scandinavian Actuarial Journal, 727 (2018) [7] N Friedman, L Cai, and X S Xie, Phys. Rev. Lett. 97, 168302 (2006) [8] V Shahrezaei and P S Swain, Proc Natl Acad Sci U S A 105, 17256 (2008) [9] I Golding, J Paulsson, S M Zawilski, and E C Cox, Cell 123, 1025 (2005) [10] A Raj and A van Oudenaarden, Cell 135, 216 (2008) [13] T Zhu, J Li, A Samanta, A Leach, and K Gall, Physical Review Letters 100, 025502 (2008) [14] P Sollich, F Lequeux, P Hebraud, and M E Cates, Physical Review Letters 78, 2020 (1997) [15] S O Rice, Bell System Technical Journal 23, 282 (1944) [16] G L´eveill´e and E Hamel, Scandinavian Actuarial Journal, 228 (1919) [17] K Borovkov and A Novikov, Statistics & Probability Letters 78, 1517 (2008) [18] O Kella and W Stadje, Journal of Applied Probability 38, 781 (2001) [44] D R Cox and V Isham, Point Processes (CRC Press, 1980) [45] J Grandell, Aspects of Risk Theory (Springer, New York, 1991) [46] D R Cox and W L Smith, Biometrika 41, 91 (1954) [47] G L´eveill´e and J Garrido, Scandinavian Actuarial Journal, 98 (2001) [48] L Tak´acs, Acta Mathematica Academiae Scientiarum Hungarica 6, 363 (1955) [49] R B Stein, Biophysical Journal 5, 173 (1965) [50] N Brunel and V Hakim, Neural Computation 11, 1621 (1999) [53] I N Burenev, First-passage properties of the jump process with a drift. The general case (2025), arXiv:2510. p. 517988 [cond-mat] [54] See Supplemental Material (SM) at https://doi. p. 5org/10.1103/bbh8-n8dt for details on analytical calculations and numerical simulations, which includes Refs. p. 5
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved AI systems can now analytically derive universal asymptotic formulas for mean first-passage time (MFPT) of renewal shot noise with arbitrary arrival statistics and exponential marks, overcoming a long-standing limitation confined to the Poisson case.
-
The improved system can predict universal scaling corrections that modulate the baseline Arrhenius law based on temporal correlations in arrivals, distinguishing between bursty and non-bursty regimes.
-
The system can provide an exact closed-form expression for all moments of renewal shot noise, which allows for a general framework to analyze extreme events in non-Markovian systems with relaxation.
-
It can characterize the full first-passage time distribution at large thresholds as exponential, providing a complete asymptotic characterization of the MFPT through the formula P(Tb > t) = exp(-t/⟨Tb⟩).
-
The system can calculate finite threshold corrections beyond the rare-event estimate ⟨TbRE⟩ in regimes where temporal correlations lead to enhanced clustering, such as when κ <= 1.
-
The improved system can provide explicit estimates for the terms Ab and Bb, allowing for quantification of finite-threshold corrections to the MFPT beyond the rare-event estimate ⟨TbRE⟩.
Sources
- Persistence and First-Passage Properties in Non-equilibrium Systems
- First-passage properties of the jump process with a drift. The general case
Related papers
- Emergent long-tail dynamics in driven magnets with dynamical frustration
- Statistical Mechanics of Dynamical System Identification
- Second law of thermodynamics in closed quantum many-body systems
- Chemical potentials from structure factors: II. Charged multi-component mixtures
- Genuine quantum scars in Floquet chaotic many-body systems
- Stochastic Thermodynamics for Autoregressive Generative Models: A Non-Markovian Perspective