Re-structure the Sage project scope and architecture

[victor-iyi]

Sage is the software stack of a much bigger project called "Enhancing Human Intelligence" which has had 3 publications so far.

• Enhancing Human Intelligence: Episode One
• Enhancing Human Intelligence: Episode Two
• Enhancing Human Intelligence: 2024 Update

---

It consists of several sub-systems including sage-ai (which is a multimodal action model with several AI agents), a decoder system that interprets brain signals into natural language, dual learning system, etc. All of these sub-system should be tightly integrated within Sage.

As the project evolves so does the scope of the Sage system. The initial goal of Sage was to be an open source knowledge graph with AI features such as Graph Neural Networks, Graph Embeddings, Semantic search etc in order to understand the world and relationships around real world objects/entities. However, there has been technological advancements that makes this approach obsolete.

---

Rather than have Sage be an Open source knowledge graph, Sage will now be a BCI system that decodes human brain signals into natural language, interprets them, processes the signals with sage-ai and perform actions based on the signals or natural language input.

Features of Sage

• Use decoded natural language/IR as input to sage-ai
• Perform actions suggested by sage-ai's response
• 🔬 Decode brain signals into natural language or intermediate representation
• 🔬 Filter response to be encoded as signals to be transmitted into the brain
• 🔬 Dual Learning System

A lot of Sage's sub-system is still in the research phase 🔬.

Such system would be built iteratively. The goal is to have sage v0.1.0 be a system that can work with natural language and perform actions based on the contents of the language using many tools that will be provided by Sage and SageAI.