Transition Geometry under Gaussian and Levy Perturbations: Time Coupling and Transition Band

Transition Geometry under Gaussian and Levy Perturbations: Time Coupling and Transition Band

by Yuzhu Shi

Formal & Physical Sciences Mathematics PBWApplied mathematicsPBWLStochastics
🎙 Yuzhu Shi 👥 2K 📅 August 12, 2026 ⏱ 43 min 👁 14 📄 original study 🧭 2026-08-15
Available in: English (current) Français

Keywords

committor functionstochastic separatrixtransition bandLevy noisemean first passage time

Summary

The seminar by Yuzhu Shi, presented at the Isaac Newton Institute, focuses on transition geometry in stochastic dynamical systems under Gaussian and Levy perturbations. The speaker begins by introducing a conceptual model of Arctic sea ice and phytoplankton biomass, which exhibits bistability between a low-biomass ‘background’ state and a high-biomass ‘productive’ state. The system is modeled with stochastic differential equations, incorporating Gaussian noise initially. Key concepts include the committor function, which gives the probability of reaching one stable state before another, and the stochastic separatrix, defined as the set where the committor equals 0.5. The speaker defines a ’transition band’ as the region between committor values 0.4 and 0.6, and shows that its width increases with noise intensity. A key result is a linear relationship between the local transition band width and noise intensity, and an inverse square relationship between the logarithm of mean first passage time and noise intensity. The talk then extends to Levy noise, which introduces jumps. The speaker discusses challenges in defining the transition band due to frequent escapes from the computational domain, and proposes a modified committor function and mean transition time. Preliminary results suggest a linear relationship between noise amplitude and local width, with the Y-direction (biomass) dominating due to a smaller stability index. The presentation concludes with open questions about numerical methods and the need for further exploration.

225 words

Critical Evaluation

Value of the Information & Strength of the Argument

The value of the information is high for researchers in stochastic dynamics and metastability. The speaker presents novel results extending previous work on Gaussian noise to Levy noise, which is more realistic for many applications. The argumentation is solid, with clear mathematical derivations and numerical simulations supporting the claims. The speaker acknowledges limitations and open questions, which adds credibility. However, the presentation is preliminary and some results are not fully validated, as the speaker notes ongoing work.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is good, with a clear mathematical framework and numerical validation. The speaker does not cite specific sources in the talk, but the work is likely based on established literature in stochastic differential equations and large deviation theory. The title accurately reflects the content. No comments were provided for analysis.

144 words

Title / Content Match

The title accurately reflects the content, focusing on transition geometry under Gaussian and Levy perturbations, with emphasis on time coupling and transition band.

Quality & Reliability

7/10

The presentation is a research seminar with clear mathematical derivations and numerical simulations. The speaker acknowledges limitations and open questions. However, the talk is preliminary and lacks peer-reviewed publication details.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

Concurring Sources

  • INI Seminar Page — The seminar is part of a programme on metastability and critical transitions, which aligns with the topic.

Contribution & Novelties

The presentation extends previous work on transition geometry under Gaussian noise to Levy noise, which is more realistic for many applications. It introduces a modified committor function and mean transition time to handle escapes from the computational domain. Preliminary results show a linear relationship between noise amplitude and local transition band width, with directional dominance depending on the stability index.

Pour aller plus loin :

  • Large deviations theory — Provides the theoretical foundation for rare event probabilities in stochastic systems.
  • Levy process — Mathematical definition and properties of Levy processes, including alpha-stable distributions.
  • Metastability — Concept of metastable states and transitions, relevant to the bistable system discussed.

107 words

Radar Profile

The radar profile shows high scores in technical level and information quantity, indicating a mathematically dense presentation with substantial content. Quality and reliability are moderate, reflecting the preliminary nature of the research. The overall balance suggests a valuable but not fully mature contribution.

Reliability 7/10