How does the Dynamic Sufficiency Gate work?

It evaluates your current toolchain (libraries like NumPy, SciPy, or languages like Julia) against the requirements of a specific GMRA operation, ensuring the agent uses the most appropriate implementation for your system.

What is 'Local Reconstructability' in this context?

It refers to the ability to reconstruct specific data cell information or visual identifiers deterministically from a seed and index, allowing different compute nodes to work on shards without a global trusted setup.

Does this skill require a paid MATLAB license?

No, the skill includes a dynamic sufficiency gate that automatically detects your environment and can fall back to GNU Octave, Julia, or a pure Python implementation if MATLAB is unavailable.

Geometric Multi-Resolution Analysis (GMRA) is a technique for hierarchical decomposition of high-dimensional data, used to approximate manifolds and discover geometric structures at different scales.

GMRA Multi-Resolution Analysis

Name: GMRA Multi-Resolution Analysis
Author: plurigrid

byplurigrid

•

Data Science & ML

Performs high-dimensional geometric multi-resolution analysis with deterministic local reconstructability across MATLAB, Julia, and Python.

The GMRA-MATLAB skill implements Geometric Multi-Resolution Analysis (GMRA) to decompose high-dimensional data into hierarchical structures using local PCA and clustering. It features a unique 'Dynamic Sufficiency' gate that evaluates environment capabilities—such as the presence of MATLAB, Octave, Julia, or Python—before execution. By leveraging deterministic coloring via Gay.jl integration, it enables local reconstructability for sharded datasets without requiring a trusted setup, making it an essential tool for topological data analysis and complex manifold learning tasks.

Key Features

018 GitHub stars

02Dynamic sufficiency gate for cross-language environment detection and fallback

03Multi-language support for MATLAB, Julia, R, and pure Python implementations

04Hierarchical data decomposition using multi-resolution geometric analysis

05Local PCA for tangent space approximation on high-dimensional manifolds

06Deterministic sharding and local reconstructability via SplitMix64 hashing

Use Cases

01Implementing sharded data processing with verifiable local reconstruction

02Bridging research between legacy MATLAB toolboxes and modern Python/Julia data stacks

03Analyzing high-dimensional datasets with non-linear manifold structures

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add plurigrid/asi gmra-matlab

For use in Claude.ai and ChatGPT

Download Skill