humemai-research 2.3.0.post2

A Machine With Human-Like Memory Systems

humemai

• Built on a cognitive architecture
  • Functions as the brain 🧠 of your own agent
  • It has human-like short-term and long-term memory
• The memory is represented as a knowledge graph
  • A graph database (JanusGraph + Cassandra) is used for persistence and fastgraph traversal
  • The user does not have to know graph query languages, e.g., Gremlin, since HumemAI handles read from / write to the database
• The interface of HumemAI is natural language, just like a chatbot.
  • This requires the Text2Graph and Graph2Text modules, which are part of HumemAI
• Everything is open-sourced, including the database

Installation

The humemai python package can already be found in the PyPI server

pip install humemai

or

pip install 'humemai[dev]'

for the development

Supports python>=3.10

Text2Graph and Graph2Text

These two modules are critical in HumemAI. At the moment, they are achieved with LLM prompting, which is not ideal. They'll be replaced with Transformer and GNN based neural networks.

Example

• example-janus-agent.ipynb: This Jupyter Notebook reads the Harry Potter book paragraph by paragraph and turns it into a knowledge graph. Text2Graph and Graph2Text are achieved with LLM prompting.
• More to come ...

Visualizaing Graph

Use JanusGraph-Visualizer to visualize the graph.

Run below:

docker run --rm -d -p 3000:3000 -p 3001:3001 --name=janusgraph-visualizer --network=host janusgraph/janusgraph-visualizer:latest

And open http://localhost:3001/ on your web browser

Work in progress

Currently this is a one-man job. Click here to see the current progress.

pdoc documentation

Click on this link to see the HTML rendered docstrings

Contributing

Contributions are what make the open source community such an amazing place to be learn, inspire, and create. Any contributions you make are greatly appreciated.

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/AmazingFeature)
3. Run make test && make style && make quality in the root repo directory, to ensure code quality.
4. Commit your Changes (git commit -m 'Add some AmazingFeature')
5. Push to the Branch (git push origin feature/AmazingFeature)
6. Open a Pull Request

Authors

• Taewoon Kim