JAX-HDC Documentation

JAX-HDC is a high-performance library for Hyperdimensional Computing (HDC) and Vector Symbolic Architectures (VSA) built on JAX.

Features

• XLA compilation and automatic kernel fusion

• Native GPU/TPU support through JAX backend

• Pure functional design enabling JAX transformations (jit, vmap, grad, pmap)

• Four VSA model implementations: BSC, MAP, HRR, FHRR

• Feature encoders for discrete, continuous, and high-dimensional data

• Classification models with test coverage

Quick Start

Installation:

git clone https://github.com/rlogger/jax-hdc.git
cd jax-hdc && pip install -e .

Basic usage:

import jax
from jax_hdc import MAP

model = MAP.create(dimensions=10000)
key = jax.random.PRNGKey(42)

x = model.random(key, (10000,))
y = model.random(key, (10000,))
bound = model.bind(x, y)
similarity = model.similarity(x, y)

Documentation Contents

User Guide

• Installation
  • Requirements
  • Installing from Source
  • Using Nix
  • GPU/TPU
• Quick Start
  • Basic Usage
  • Classification Pipeline
  • Key Concepts
• HDC Learning
  • Configuration and Imports
  • Datasets
  • Encoding
  • Training
  • Testing
  • Running the Full Example
• Examples
  • Basic Operations
  • Classification
  • Kanerva’s “Dollar of Mexico”

API Reference

• API Reference
  • bind_bsc()
  • bundle_bsc()
  • inverse_bsc()
  • hamming_similarity()
  • bind_map()
  • bundle_map()
  • inverse_map()
  • cosine_similarity()
  • permute()
  • cleanup()
  • BSC
  • BSBC
  • MAP
  • HRR
  • FHRR
  • CGR
  • MCR
  • VTB
  • RandomEncoder
  • LevelEncoder
  • ProjectionEncoder
  • KernelEncoder
  • GraphEncoder
  • CentroidClassifier
  • AdaptiveHDC
  • LVQClassifier
  • RegularizedLSClassifier
  • SparseDistributedMemory
  • HopfieldMemory
  • AttentionMemory
• Functional Module
  • Binary Spatter Code Operations
  • MAP Operations
  • HRR Operations
  • Universal Operations
  • Batch Operations
  • Example Usage
• VSA Models
  • Base Class
  • Binary Spatter Codes
  • Multiply-Add-Permute
  • Holographic Reduced Representations
  • Fourier HRR
  • Factory Function
  • Example Usage
• Embeddings Module
  • RandomEncoder
  • LevelEncoder
  • ProjectionEncoder
  • KernelEncoder
  • GraphEncoder
• Models Module
  • CentroidClassifier
  • LVQClassifier
  • RegularizedLSClassifier
  • AdaptiveHDC
• Memory Module
  • SparseDistributedMemory
  • HopfieldMemory
• Utilities Module
  • Configuration
  • Random Number Generation
  • Benchmarking
  • Validation
  • Helpers
  • Example Usage

Additional Information

• Contributing

Indices and tables

• Index

• Module Index

• Search Page