Create corpus for NL-to-MeTTa LLM

Longer description

Context and background:

SingularityNET Foundation, in collaboration with other partners such as the OpenCog Foundation and TrueAGI, is working toward a scalable implementation of the Hyperon AGI framework running on decentralized infrastructure, and toward the implementation of the PRIMUS cognitive architecture within this framework.

Hyperon and PRIMUS are complex systems involving multiple components, which need to demonstrate appropriate functionalities both individually and in combination. This RFP aims to address a portion of this overall need by funding the initial iteration of a significant component of PRIMUS within Hyperon: the development of a comprehensive corpus for NL-to-MeTTa language model training.

The purpose of this corpus is to enable the creation of an AI-powered MeTTa coding assistant, which will assist users in generating correct and functional MeTTa code. This coding assistant will play a crucial role in lowering the barrier to entry for MeTTa, thereby accelerating the broader AGI development within the Hyperon framework.

MeTTa (Meta Type Talk) is a multi-paradigm language for declarative and functional computations over knowledge metagraphs, designed specifically to meet the needs of Artificial General Intelligence (AGI). It is an innovative and relatively new language, and might sometimes come with a learning curve for starters. While there are also plenty of resources and tutorials that one can reference for his/her needs, they may not be enough to fully cover all possible unique usages. Therefore it will be beneficial to have a coding assistant for MeTTa that can instantly help people to create what they want with MeTTa.

There are materials like the official documentation, tutorials, and repositories on GitHub that contain MeTTa programs, which include those that are created and written by our community members. All these are valuable resources that can be used for the purpose of creating such a MeTTa coding assistant, but they are scattered around and often not in a ready-to-use format, besides the volume of these resources in total may not be enough to even fine-tune an AI model given how new the MeTTa language is at the moment. Therefore the purpose of this RFP is to address this need by:

Identifying and converting from the existing MeTTa resources to a format that is ready-to-use for the purpose of creating a MeTTa coding assistant

Generate/synthesize new data in the same format for the same purpose

Collaboration

This RFP will be followed by subsequent RFPs that make use of the MeTTa corpus to train or fine-tune an AI model as a coding assistant for MeTTa.

RFP expected outcomes:

• MeTTa corpus
  • A MeTTa corpus that can be used to train or fine-tune an AI model as a MeTTa coding assistant
• OSS code
  • All the code that was used to create the MeTTa corpus that others can also run and replicate the same corpus creation process
• Thorough documentation
  • Provide comprehensive documentation detailing how the MeTTa corpus is gathered and/or generated/synthesized