Reference implementation of "An Algorithm for Routing Vectors in Sequences" (Heinsen, 2022) and "An Algorithm for Routing Capsules in All Domains" (Heinsen, 2019), for composing deep neural networks.

Implementation of deep implicit attention in PyTorch

Holistic system for storage and transformation of information based on associative model of data. Целостная система для хранения и обработки информации, основанная на ассоциативной модели данных.

Physics-inspired transformer modules based on mean-field dynamics of vector-spin models in JAX

Cognitive Computing with Associative Memory

This repository contains the official code for Energy Transformer---an efficient Energy-based Transformer variant for graph classification

Implementation of Hopfield Neural Network in Python based on Hebbian Learning Algorithm

[ICML 2024] Uniform Memory Retrieval with Larger Capacity for Modern Hopfield Models

Recurrent predictive coding networks for associative memory employing covariance learning

Synapse Context Engine (SCE) is a brain-inspired hypergraph-based AI memory architecture for persistent context, coherent reasoning, and long-term memory, designed for transparency and safety.

A TypeScript-based MCP server implementing a memory system inspired by neurodivergent cognitive styles. It organizes thoughts into five districts (knowledge domains), ranks search results using BM25 semantic ranking, and stores memories as a persistent knowledge graph with bidirectional connections.

Holographic vectors you can compute with. Bind structure, bundle sets, unbind components cross NumPy, PyTorch, and JAX.

PyTorch implementation of IJCAI 2020 paper Exploiting Neuron and Synapse Filter Dynamics in Spatial Temporal Learning of Deep Spiking Neural Network

Minimum Description Length Hopfield Networks

Open Source C++ Library for Pseudo-inverse Fully Connected Recurrent Neural Networks (from my PhD)

Metacognitive AI architecture combining Modern Hopfield Networks with entropy-based confusion detection for insight-like problem solving. Implements energy-based search, adaptive beta scaling, and a Cognitive Database integration for analogical reasoning and topology tunneling.

ELEC 8330: Computational Intelligence [CI] | Semester IV | MEng Computer Engineering

Source-codes and examples of quaternion-valued recurrent projection neural networks on unit quaternions

Publications by Peter Overmann

AI agent memory using Modern Hopfield Networks — no LLM calls, no database, one matrix multiply. MCP server for Cursor, Claude Code, and other AI coding agents.