ICM 2026 Plenary Lecture - Horng-Tzer Yau

ICM 2026 Plenary Lecture - Horng-Tzer Yau

Formal & Physical Sciences Mathematics PBMathematics
🎙 Horng-Tzer Yau 👥 58K 📅 August 17, 2026 ⏱ 53 min 👁 13 📄 original study 🧭 2026-08-17
Available in: English (current) Français

Keywords

random matricesWigner universalityAnderson localizationquantum diffusiondelocalization

Summary

Horng-Tzer Yau delivers a plenary lecture at ICM 2026 on random matrices, focusing on Wigner universality, Anderson localization, and recent advances. He begins by introducing the Anderson tight-binding model and the concept of localization, then contrasts it with Wigner’s vision of universal spectral statistics for complex systems. The lecture connects these ideas through random band matrices, which interpolate between the two extremes. Yau outlines the three-step strategy for proving universality: a priori estimates on Green’s functions, universality for matrices with small Gaussian noise (Dyson’s conjecture), and a comparison argument. He emphasizes the role of quantum diffusion and quantum unique ergodicity in extending results to band matrices. Recent work proves delocalization and universality for band matrices in dimensions 1, 2, and 3 under certain bandwidth conditions, and establishes quantum diffusion for the Anderson model. The lecture also introduces a new class of integrable fixed points called primitive hierarchies, which serve as approximations for delocalization problems. The talk concludes with results on regular graphs and the Tracy-Widom fluctuations.

166 words

Critical Evaluation

Value of the Information & Strength of the Argument

The lecture provides a comprehensive overview of recent breakthroughs in random matrix theory, with a strong emphasis on rigorous proofs and the underlying strategies. Yau clearly explains the three-step strategy and how it applies to both Wigner and band matrices, highlighting the key challenges and how they were overcome. The argumentation is solid, based on published results and collaborations with other experts. The introduction of primitive hierarchies as new integrable fixed points is a novel contribution that offers a fresh perspective on delocalization problems. The lecture is dense but well-structured, making it valuable for researchers and advanced students.

Scientific Rigor, Source Quality, Title Accuracy

The lecture is scientifically rigorous, presenting results that have been peer-reviewed and published in top journals. Yau references numerous collaborators and prior works, though specific citations are not explicitly listed in the video. The title accurately reflects the content, covering Wigner universality, Anderson localization, and beyond. The lecture is part of the ICM 2026 plenary series, indicating high academic standing. No comments were provided, so no analysis of public reception is possible.

185 words

Title / Content Match

The title accurately reflects the content: a plenary lecture on random matrices, covering Wigner universality, Anderson localization, and related topics.

Quality & Reliability

9/10

Lecture by a leading mathematician (Fields Medal-level contributions) presenting rigorous results with proofs and references to published work. High reliability, though not peer-reviewed in this format.

Key Moments

Cited Sources

Concurring Sources

Contribution & Novelties

The lecture presents recent breakthroughs in random matrix theory, including the resolution of the Wigner-Dyson conjecture, delocalization and universality for random band matrices, and the introduction of primitive hierarchies as new integrable fixed points. It also establishes quantum diffusion for the Anderson model. These results significantly advance the understanding of spectral statistics in complex systems.

Pour aller plus loin :

98 words

Radar Profile

The radar profile shows very high scores in information quality, technical level, and reliability, with slightly lower but still high scores in information quantity and global reliability. This indicates a dense, rigorous, and highly technical lecture suitable for experts.

Reliability 9/10