Dracon Intelligence System
Current Status: Under Development
The Dracon Intelligence System (D.I.S) is an ongoing exploration into the use of generative AI models.
Previously, this has focused on using neural networks for semantic analysis with rigid responses.
This approach has altered in recent days with the release of ChatGPT and associated models.
The approach currently is:
1. Speech recognition using OpenAI's Whisper model.
2. Semantic analysis using a RASA conversational AI model.
3. Semantic action using a C# framework
4. Generate response using ChatGPT 3.5-turbo
5. Text to speech using (?) model
Given the number of moving parts here, there are various element all in construction.
The window below was being used as our communications framework, but instead of an entirely server based approach, I'm focusing on a capable client this time, which can update user data to a centralised database, instead of a server authoritative approach.
The chat below wont work, but I will let you all know when it does!
Peter Carey
Previously, this has focused on using neural networks for semantic analysis with rigid responses.
This approach has altered in recent days with the release of ChatGPT and associated models.
The approach currently is:
1. Speech recognition using OpenAI's Whisper model.
2. Semantic analysis using a RASA conversational AI model.
3. Semantic action using a C# framework
4. Generate response using ChatGPT 3.5-turbo
5. Text to speech using (?) model
Given the number of moving parts here, there are various element all in construction.
The window below was being used as our communications framework, but instead of an entirely server based approach, I'm focusing on a capable client this time, which can update user data to a centralised database, instead of a server authoritative approach.
The chat below wont work, but I will let you all know when it does!
Peter Carey