This repository provides a comprehensive survey of memory research in Artificial Intelligence, targeting researchers and practitioners in the field. It offers a curated list of papers, datasets, methods, and tools, aiming to systematize the understanding and advancement of AI memory mechanisms.

How It Works

The repository categorizes memory research into key areas: Taxonomy, Operations, Topics (Long Term, Long Context, Parametric, Multi-source), Datasets, Methods, and Tools. It systematically lists and categorizes relevant academic papers, datasets for evaluation, and software tools used in memory research, providing a structured overview of the landscape.

Quick Start & Requirements

This repository is a curated list and does not require installation or execution. It serves as a reference guide.

Highlighted Details

• Extensive categorization of memory research into taxonomy, operations, and specific topics.
• Detailed tables of datasets for evaluating long-term, long-context, parametric, and multi-source memory.
• Comprehensive lists of component-level, framework-level, and application-layer tools for memory management.
• Comparative analysis of human vs. AI memory mechanisms.

Maintenance & Community

The repository is actively maintained, with recent updates and contributions welcomed. It cites a specific paper for reference: "Rethinking Memory in AI: Taxonomy, Operations, Topics, and Future Directions."

Licensing & Compatibility

The repository is licensed under the MIT License, permitting commercial use and integration into closed-source projects.

Limitations & Caveats

As a survey and reference repository, it does not provide executable code or implementations. The content is a curated list of external resources, and the quality and applicability of those resources depend on the original sources.