diagram.md

graph TD
    UI["User Interface (Chat Interface)"] --> PP["Processing Pipeline (Preprocessing, Context Management, NER Extraction)"]
    PP --> RG["Response Generation (GPT Model + LoRA)"]
    PP --> RK["Retrieval Knowledge Store (FAISS Index)"]
    RG --> FBC["Feedback Collection (Explicit/Implicit)"]
    RG --> TKS["Training Knowledge Store (User Inputs, Responses, Feedback, Metadata)"]
    FBC --> ATP["Asynchronous Training Pipeline (Fine-Tuning with Backpropagation)"]
    RK --> ATP

Loading

• Assistant needs both knowledge store and training pipeline.
• Training pipeline is to provide better prediction on responses.
• Training pipeline utilizes knowledge store and a pretrained model to generate a DPO example from a user query and an assistant response.
• The DPO example is then used to fine-tune the model.
• The knowledge store is used to store the knowledge that the assistant has learned.
• The knowledge store is updated with new knowledge when the assistant learns something new.
• The knowledge store is compressed using FAISS to store the knowledge in a more efficient way.