GitHub - harishsg993010/HawkinsDB: HawkinsDB is our take on giving AI systems a more human-like way to store and recall information, inspired by how our own brains work. Based on Jeff Hawkins' Thousand Brains Theory
Dec 27, 2024 - github.com
HawkinsDB is a neuroscience-inspired memory layer designed to enhance large language model (LLM) applications by providing a more human-like way to store and recall information. Inspired by Jeff Hawkins' Thousand Brains Theory, it integrates semantic, episodic, and procedural memory into a unified framework, allowing AI systems to manage complex information with context-aware queries. Unlike traditional vector databases, HawkinsDB uses concepts like Reference Frames and Cortical Columns to create a robust, adaptable system that offers transparency in understanding data connections and decision-making processes.
The platform supports Python 3.10 or higher and requires an OpenAI API key for LLM operations, with storage options including SQLite and JSON. It offers smart integrations like ConceptNet for knowledge enrichment and relationship discovery. HawkinsDB is actively developed with a focus on enhanced multi-modal processing, performance optimization, and extended LLM provider support. It is available under the MIT License, and contributions are encouraged through a structured development process.
Key takeaways:
HawkinsDB is a neuroscience-inspired memory layer for AI systems, designed to store and recall information in a human-like way.
It unifies semantic, episodic, and procedural memory into a single framework, enabling context-aware queries and understanding of data relationships.
The system is based on concepts like Reference Frames and Cortical Columns, allowing for multi-perspective information processing.
HawkinsDB supports integrations like ConceptNet for knowledge enrichment and offers storage options like SQLite and JSON.
