ICM 2026 Invited Section Lecture - Patrick Farrell

Value of the Information & Strength of the Argument

The value of the information presented is high: it addresses a fundamental problem in computational science — the risk of incomplete solution sets — and proposes a practical, mathematically justified remedy. The argumentation is particularly strong because the speaker builds each step from a clear motivational example (the metamaterial), identifies the failure modes of classical methods, and then constructs the deflation formalism almost from first principles. The advantages and limitations are both explicitly stated: the method does not guarantee global convergence, but it is strictly better than having only one solution. The use of concrete bifurcation diagrams, experimental footage, and quantitative statements (e.g., rank-one update of the Jacobian) makes the claims easy to grasp and verify. The comparison with classical numerical bifurcation analysis is honest and well balanced, acknowledging its power while pointing out practical obstacles.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is unmistakable: the talk integrates decades of numerical analysis literature (Wilkinson’s polynomial deflation, Brown & Gearhart’s general maps, Keller’s bifurcation analysis) and the speaker’s own peer-reviewed contributions (the +1 shifted deflation in 2015, the infinite-dimensional theory from 2013). Each key step is justified with mathematical reasoning, not asserted. The title faithfully reflects the content: this is indeed an ICM invited section lecture on computing multiple solutions via deflation, delivered at a specialized mathematical level.

The title accurately reflects the content: an invited section lecture on computing multiple solutions of nonlinear equations using deflation.

9/10

Lecture by a recognized researcher in numerical analysis, presenting peer-reviewed results. The mathematical arguments are rigorous, placed in a well-documented historical context (Wilkinson, Brown & Gearhart, Keller, Farrell et al.), and limitations are honestly addressed.

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

• 01:30Motivating example: mechanical metamaterial with negative Poisson ratio
• 04:00Experimental demonstration of two stable configurations in a metamaterial
• 07:00Mathematical problem formulation: f(u, λ) = 0
• 10:00Classical numerical bifurcation analysis (Herb Keller, 1970s)
• 15:00Limitations: disconnected branches and auxiliary problems
• 20:00Newton's method review and introduction of deflation
• 26:00History: Wilkinson, Brown & Gearhart (1971), shifted deflation (+1) by Farrell et al. (2015)
• 33:00Efficient implementation using Sherman-Morrison formula
• 37:00Examples: Yamabe equation from Riemannian geometry

• Computing Multiple Solutions of Systems of Nonlinear Equations with Deflation (abstract) — Official abstract of the talk posted in the video description.
• Brown & Gearhart (1971) — deflation operators for nonlinear maps — Mentioned around 27 minutes as the original source of the deflation concept for maps R^n to R^n.
• Wilkinson — polynomial deflation — Mentioned as the precursor of the deflation idea in the context of polynomials.
• Keller (1970s) — numerical bifurcation analysis — Mentioned as the pioneer of classical numerical bifurcation analysis.
• Farrell et al. (2015) — shifted deflation operator — The speaker's main contribution: adding '+1' to the deflation operator, presented around 29 minutes.

• Newton's method — Adjacent foundational method; the base algorithm that deflation modifies.
• Numerical continuation — The classical approach to which deflation is presented as complementary.
• Sherman–Morrison formula — Supports the claimed efficiency of the deflated Newton step computation.

• 1990 textbook on numerical bifurcation analysis (quoted in the talk) — The speaker quotes this book as claiming deflation is unreliable ('often a matter of seeming chance whether one obtains an additional solution'), a criticism that motivated the shifted deflation operator.