Persistence exponents of self-interacting random walks

arXiv:2410.18699 · cond-mat.stat-mech, math.PR · Submitted 2024-10-24 · Read on arXiv

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: I'm Kai.

Mira: And I'm Mira, here with Lev.

Kai: We're looking at the paper: "Persistence exponents of self-interacting random walks".

Paper summary: Kai: So, looking at the whole paper "Persistence exponents of self-interacting random walks," the main thing is that Ray-Knight theory lets you get exact values for the splitting probability and these persistence exponents for SIRWs >

Mira: It’s interesting because they say that this arcsine law result for PSRW gamma and SESRW kappa beta is new, which is a specific finding they are highlighting >

Lev: But there’s a limitation mentioned in the paper, so remember that expressions involving Iz do not hold for all non-Markovian processes, but they are valid specifically for SIRWs >

Kai: So what does this mean for us who only listen to the show? It means we can trust these exact numbers when we apply them to models of transport where interactions matter, like in complex fluids or disordered materials >

Mira: It shows that even with long-range memory effects from the walker interacting with its own past path, we can nail down the long-time decay behavior precisely for these specific walk types >

Lev: For hardware development, this means if you are building an error correction code based on these models, you have a precise value for how fast things decay before they hit the absorbing state >

Kai: This paper provides solid benchmarks for future analysis of more realistic self-interacting processes that might not fit neatly into these three classes >

Conclusion: Kai: So we've been looking at how these self-interacting random walks behave in the long run, and this paper, "Persistence exponents of self-interacting random walks," gives us some exact numbers for those decay rates.

Mira: Exactly, it’s about getting those persistence exponents for three different types of walks—the once-reinforced one, the polynomially self-repelling one, and the sub-exponential ones.

Lev: The authors used Ray Knight theory to calculate these exact values for the splitting probability and then derived the persistence exponent from that. That's a pretty heavy piece of math to get right.

Kai: I mean, getting those exact formulas is important because we can actually test them against what we build in the lab, you know? We can check if our cooling or our error correction actually matches these theoretical predictions.

Mira: Right, and the authors point out that this arcsine law result for those two walk types is something new. That’s a real piece of new information about how these systems behave when they interact.

Lev: From an engineering standpoint, those exact persistence exponents mean we have a better benchmark for designing error correction codes for these complex transport processes. It shows us the fundamental limits of how fast information can decay in these interacting systems before it hits an absorbing state.

Kai: So what this really means is that we’ve moved past just having approximations and got concrete values that we can use to design better hardware, whether that's for qubits or something else involving these walks.

Mira: It confirms the power of Ray Knight theory in giving us those exact results for self-interacting random walks, which are fundamental models for transport.

Lev: And it sets a clear path forward for analyzing more complicated, real-world models where the interactions might not fit neatly into those three categories they studied.

J. Brémont, –L. R⠼gnier, R. Voituriez, O. B⡎nichou

Laboratoire de Physique Théorique de la Matiére Condenséee, CNRS/Sorbonne Universit⠼ e

cond-mat.stat-mech, math.PR

Submitted: 2024-10-24

Updated: 2024-10-24

Journal ref: PhysRevLett.134.197103 (2025)

DOI: 10.1103/PhysRevLett.134.197103

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 80/100

The gist: The persistence exponent, which characterises the long-time decay of the survival probability of stochastic processes in the presence of an absorbing target…

Key concepts

Persistence Exponent (θ)
This exponent describes how long the survival probability of a stochastic process decays over time when an absorbing target is present. It characterizes the long-time behavior of the system's survival, following a power-law decay S(x, t) ∝ t − θ.
Self-Interacting Random Walks (SIRWs)
These are fundamental models of transport where the movement of a particle depends on its previous positions. The paper focuses on three specific universality classes: once-reinforced (SATWϕ), polynomially self-repelling (PSRWγ), and sub-exponential self-repelling (SESRWκ,β).
Splitting Probability
This probability relates to the distribution of how a walk splits into two paths. The authors use Ray-Knight theory to compute this probability, which is essential for deriving the persistence exponent through scaling limits.

Terminology

Summary

The persistence exponent, which characterises the long-time decay of the survival probability of stochastic processes in the presence of an absorbing target, plays a key role in quantifying the dynamics of fluctuating systems. p. 0 This Letter fills this gap by providing exact values of θ for the broad class of self-interacting random walks (SIRWs), which are fundamental models of transport. p. 0

Introduction to SIRW and Persistence Exponent

The persistence exponent, denoted as θ, characterizes the long-time decay of the survival probability S(x, t) following a power-law decay of the form S(x, t) ∝ t −θ. While exact computation of S(x, t) is usually out of reach for non-Markovian processes, this paper provides exact values for θ across three universality classes of SIRWs. p. 0 These classes are defined by the large-scale behavior of the walk as n → ∞. p. 1 The three classes are the once-reinforced random walk (SATWϕ), the polynomially self-repelling walk (PSRWγ), and the sub-exponential self-repelling walk (SESRWκ,β). p. 1

Determination of Splitting Probability

The authors employ a two-step strategy to obtain the persistence exponent, first computing the splitting probability Q+(k, −m) using Ray-Knight theory. The splitting probability is related to the observable q+(k, L), which is defined as P(Ik = [−(L − k) + 1, k]). p. 1 This distribution q+(k, L) is then related to the edge local times (LTk(x)) via the identity Q+(k, −m) = k+Xm−1j=k q+(k, j). p. 2 The key identity realizing this is q+(k, L) = X∞ a=0 P(Ξ0 = L − kLTk(0) = a)P(LTk(0) = a). p. 2

Scaling Limits and Universal Results

The computation of q+(z), the scaling limit of q+(k, L), leads to the exact expression for the splitting probability Q+(z). This result shows that for SATWϕ, Q+(z) = I1−z(ϕ), and for PSRWγ and SESRWκ,β, Q+(z) involves an arcsine law. p. 3 The persistence exponent θ is then derived from the scaling limit of Q+(z) using the relation Q+(z) ∝ z→1 (1 − z) Φ(dwθ). p. 1

Persistence Exponent Results

The exact determination of the persistence exponent θ for each class is given by the formula θ =  ϕ2 (SATWϕ) 1/4 (PSRWγ) 1/2 κ+1 κ+2 (SESRWκ,β). This result is exact and confirmed numerically in FIG. p. 3 4. The results provide benchmarks for the future analysis of more realistic models of self-interacting processes. p. 3

Conclusion and Implications

Ray-Knight theory provides an exact determination of both the splitting probability and the persistence exponent for SIRWs. The arcsine law for PSRWγ and SESRWκ,β is new to the best of our knowledge. p. 3 This demonstrates that expressions involving Iz do not hold for all non-Markovian processes, but are valid for SIRWs. p. 3 The persistence exponent from the splitting probability yields the exact results presented in Eq. p. 3 (2). The paper concludes by providing exact results for the broad class of SIRWs. p. 0

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

  1. Exact Persistence Exponent Determination for Self-Interacting Random Walks The improved system can exactly determine theta for all physically relevant SIRWs by leveraging Ray-Knight theory to compute splitting probabilities and then using a general scaling relation relating persistence exponents to local behaviors of splitting probabilities, yielding the result theta = 1/2 (for Markov processes), 1/4 (for PSRW gamma), and 1/(2 kappa + 1) (for SESRW kappa, beta).

  2. Quantitative Characterization of Exploration Processes The system can provide a quantitative characterization of the exploration process of SIRWs by determining the splitting probability Q+(k, -m) for these processes. This allows for a precise quantification of the likelihood of specific outcomes in non-Markovian processes, extending beyond simple examples like the Gambler’s ruin problem.

  3. Universal Splitting Probability Formula The system can derive an exact expression for the splitting probability Q+(z) = 1/(1-z) phi (SATW phi), 1/2 for PSRW gamma, and 2 pi arcsin sqrt(1 - z) for SESRW kappa, beta. This universal form is new to the best of our knowledge and shows that expressions involving the regularised incomplete Beta function Iz are involved in splitting probabilities for a whole class of strongly non-Markovian processes, namely SIRWs.

Sources

Related papers