ICM 2026 Special Section Lecture - Remco Van der Hofstad

Value of the Information & Strength of the Argument

The value of the lecture is exceptionally high. It synthesizes recent research (including the speaker’s own work with collaborators) and places it within a historical context by crediting Yule, Simon, and Barabási-Albert. The argumentation is robust, building logically from definitions of random graph models to the identification of critical thresholds and then to the nuanced critical behavior. The ‘power to the masses’ vs ‘power to the wealthy’ dichotomy is a compelling and original conceptual framework that highlights the profound impact of heavy-tailed distributions on network phenomena. The mathematical reasoning is solid, clearly explaining the role of second and third moments in determining whether the exploration process has finite variance, which dictates the scaling behavior and the concentration of the giant component. The arguments not only present proven results but also motivate their significance, leaving a clear understanding of the underlying mechanisms.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is exemplary. The lecture explicitly references seminal works, including Yule (1925), Simon (1955), Barabási-Albert (1999), and Aldous (1997), demonstrating careful attention to intellectual history and foundational theory. The quality of sources is high, grounding the talk in established probability and statistical physics literature. The speaker clearly delineates between results that have been rigorously proven and those that remain open or depend on specific conditions. The title (ICM 2026 Special Section Lecture) is an official designation, while the description provides the precise scientific topic; the content aligns exactly with this, as the lecturer focuses exclusively on percolation and its surprises in random graphs. No comments were provided for analysis.

The official title refers to the lecture designation at ICM 2026, while the description accurately specifies the actual topic: 'Surprises in Percolation on Random Graphs'. The content fully matches the described topic.

9/10

Presentation by a leading expert in applied probability, based on recent peer-reviewed research and seminal historical papers. The lecture is rigorous, mathematically precise, and clearly distinguishes proven theorems from conjectures and open questions.

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

• 00:00Introduction and overview of complex networks and network science.
• 03:00Introduction of the scale-free paradigm and log-log plots for degree distributions.
• 07:00Presentation of the preferential attachment model, referencing Yule, Simon, and Barabási-Albert.
• 12:00Introduction of the Poisson random graph and the definition of vertex weights.
• 18:00Discussion of percolation on finite graphs and the concept of the critical window.
• 22:00First theorem: critical value identification for preferential attachment, with the surprise that pc = 0 for a certain regime.
• 26:00Second theorem: critical value for Poisson random graph equal to 1/mu, again showing pc = 0 under certain conditions.
• 30:00Introduction of the exploration process and the seminal work of Aldous (1997) on critical components.
• 35:00Surprising critical behavior: 'power to the masses' with finite third moment vs 'power to the wealthy' with infinite third moment.
• 40:00Summary of key results and concluding remarks on the 'surprises' in percolation on random graphs.

• Yule (1925) - A Mathematical Theory of Evolution — Early introduction of the preferential attachment mechanism.
• Simon (1955) - On a class of skew distribution functions — Rediscovery of the Yule process due to lack of stochastic process theory at Yule's time.
• Barabási & Albert (1999) - Emergence of scaling in random networks — Extremely influential paper that linked large real-world datasets to growth mechanisms and power-law degrees.
• Aldous (1997) - Brownian excursions, critical random graphs and the multiplicative coalescent — Seminal paper establishing the exploration process for studying the critical behavior of random graphs.

• Barabási & Albert (1999) - Emergence of scaling in random networks — Supports the scale-free paradigm and the importance of the preferential attachment model as the foundation for the lecture's subject.
• Aldous (1997) - Brownian excursions, critical random graphs and the multiplicative coalescent — Supports the methodological approach to studying critical components via exploration processes, which is central to the later part of the talk.