Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

\chisao{}: A GPU-Native Parallel Optimizer for Multimodal Black-Box Functions via Convergence-Anticonvergence Oscillation

About

Finding all modes of a multimodal black-box function is a fundamental challenge in optimization, Bayesian inference, and scientific computing. Existing approaches -- basin-hopping, CMA-ES, multistart gradient descent -- operate sequentially and cannot exploit the massive parallelism of modern GPU hardware. We introduce \chisao{} (\textbf{C}onvergence-\textbf{H}alt-\textbf{I}nvert-\textbf{S}tick-\textbf{A}nd-\textbf{O}scillate), a GPU-native population optimizer that runs an entire sample batch simultaneously and exploits a deliberate convergence-anticonvergence oscillation cycle to escape local traps while freezing confirmed modes. The structural move is asymmetric: samples that reach true peaks are frozen (``stuck'') and preserved, while the rest keep exploring via momentum-based anti-convergence and stochastically smoothed gradients. Adaptive reseeding via two complementary strategies (Repulse Monkey and Golden Rooster) maintains population diversity throughout. On all 42 functions of the Simon Fraser University optimization benchmark suite across dimensions $d \in \{2, 4, 8, 16, 32, 64\}$, \chisao{} achieves \textbf{100\%} mode recovery where all CPU baselines collapse at $d \geq 8$ on the hardest multimodal functions, at up to \textbf{$34\times$} speedup over basin-hopping on functions where all methods succeed (Michalewicz $d=64$) and up to \textbf{$39\times$} on unimodal functions (Rotated Hyper-Ellipsoid $d=64$, pure GPU dividend). All benchmarks evaluate the objective by value alone -- gradients come from finite differences -- so the reported speedups are a derivative-free worst case. Under substantial likelihood noise ($\sigma_{\mathrm{noise}}$ up to 1.0), mode detection remains 100\% reliable. The algorithm is available as a standalone open-source Python package on PyPI.

Ira Wolfson• 2026

Related benchmarks

TaskDatasetResultRank
Mode recoveryDixon-Price Group B (d=8)
Mode Recovery Rate100
10
Global OptimizationAckley d = 2
Mean Wall-Clock Time (s)3.7
5
Global OptimizationLevy d=8
Mean Time (s)17.5
5
Global OptimizationMichalewicz d=8
Mean Wall-Clock Time (s)7.4
5
Global OptimizationMichalewicz d=32
Mean Time (s)12
5
Global OptimizationMichalewicz d=64
Mean Wall-Clock Time (s)23.8
5
Global OptimizationRot. Hyper-Ellipsoid (d=2)
Mean Wall-Clock Time (s)0.7
5
Global OptimizationRot. Hyper-Ellipsoid d=8
Mean Wall-Clock Time (s)4.7
5
Mode recoveryGroup C functions fixed-2D multimodal
Cross-in-Tray Success Rate100
5
Mode recoverySphere Group B (d=2)
Mode Recovery Rate100
5
Showing 10 of 55 rows

Other info

Follow for update