MétaCan
Menu
Back to cohort
Record W4311941380 · doi:10.1137/21m1449579

Linear Asymptotic Convergence of Anderson Acceleration: Fixed-Point Analysis

2022· article· en· W4311941380 on OpenAlex

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

affAt least one author lists a Canadian institution in the pinned OpenAlex snapshot.
fundA Canadian funder is recorded on the work.

Bibliographic record

VenueSIAM Journal on Matrix Analysis and Applications · 2022
Typearticle
Languageen
FieldComputer Science
TopicMatrix Theory and Algorithms
Canadian institutionsUniversity of Waterloo
FundersNatural Sciences and Engineering Research Council of Canada
KeywordsMathematicsFixed pointLipschitz continuityDifferentiable functionAccelerationMathematical analysisConvergence (economics)Applied mathematicsFunction (biology)Weak convergencePhysicsComputer science

Abstract

fetched live from OpenAlex

.We study the asymptotic convergence of AA( \(m\) ), i.e., Anderson acceleration (AA) with window size \(m\) for accelerating fixed-point methods \(x_{k+1}=q(x_{k})\) , \(x_k \in \mathbb{R}^n\) . Convergence acceleration by AA( \(m\) ) has been widely observed but is not well understood. We consider the case where the fixed-point iteration function \(q(x)\) is differentiable and the convergence of the fixed-point method itself is root-linear. We identify numerically several conspicuous properties of AA( \(m\) ) convergence: First, AA( \(m\) ) sequences \(\{x_k\}\) converge root-linearly, but the root-linear convergence factor depends strongly on the initial condition. Second, the AA( \(m\) ) acceleration coefficients \(\boldsymbol{\beta }^{(k)}\) do not converge but oscillate as \(\{x_k\}\) converges to \(x^{*}\) . To shed light on these observations, we write the AA( \(m\) ) iteration as an augmented fixed-point iteration \(\boldsymbol{z}_{k+1} =\Psi (\boldsymbol{z}_k)\) , \(\boldsymbol{z}_k \in \mathbb{R}^{n(m+1)}\) , and analyze the continuity and differentiability properties of \(\Psi (\boldsymbol{z})\) and \(\boldsymbol{\beta }(\boldsymbol{z})\) . We find that the vector of acceleration coefficients \(\boldsymbol{\beta }(\boldsymbol{z})\) is not continuous at the fixed point \(\boldsymbol{z}^{*}\) . However, we show that, despite the discontinuity of \(\boldsymbol{\beta }(\boldsymbol{z})\) , the iteration function \(\Psi (\boldsymbol{z})\) is Lipschitz continuous and directionally differentiable at \(\boldsymbol{z}^{*}\) for AA(1), and we generalize this to AA( \(m\) ) with \(m\gt 1\) for most cases. Furthermore, we find that \(\Psi (\boldsymbol{z})\) is not differentiable at \(\boldsymbol{z}^{*}\) . We then discuss how these theoretical findings relate to the observed convergence behavior of AA( \(m\) ). The discontinuity of \(\boldsymbol{\beta }(\boldsymbol{z})\) at \(\boldsymbol{z}^{*}\) allows \(\boldsymbol{\beta }^{(k)}\) to oscillate as \(\{x_k\}\) converges to \(x^{*}\) , and the nondifferentiability of \(\Psi (\boldsymbol{z})\) allows AA( \(m\) ) sequences to converge with root-linear convergence factors that strongly depend on the initial condition. Additional numerical results illustrate our findings for several linear and nonlinear fixed-point iterations \(x_{k+1}=q(x_{k})\) and for various values of the window size \(m\) .KeywordsAnderson accelerationfixed-point methodroot-linear convergenceasymptotic convergence factorMSC codes65B0565F1065H1065K10

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.001
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.974
Threshold uncertainty score0.774

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0010.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0010.005
Science and technology studies0.0010.000
Scholarly communication0.0000.000
Open science0.0010.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0010.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.011
GPT teacher head0.273
Teacher spread0.262 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it