Upgrade to Pro — share decks privately, control downloads, hide ads and more …

B.A.A.S - BeTask AI and Automation for SMEs

B.A.A.S - BeTask AI and Automation for SMEs

B.A.A.S - BeTask AI and Automation for SMEs

LINE Developers Thailand

October 30, 2024
Tweet

More Decks by LINE Developers Thailand

Other Decks in Technology

Transcript

  1. Transform into embedding Results Flow Overview Inform Owner Embedding to

    Vector Query Vector Generate Image Narrowcast
  2. Transform into embedding Results Inform Owner Embedding to Vector Query

    Vector Generate Image Narrowcast CHECK POINT # 1
  3. Store Vector - Create Database • name : The name

    of the index • dimension : The dimensions of the vectors to be inserted in the index. • spec : The spec object defines how the index should be deployed. • metric : The distance metric to be used for similarity search. You can use 'euclidean', 'cosine', or 'dotproduct'.
  4. Transform into embedding Results Inform Owner Embedding to Vector Query

    Vector Generate Image Narrowcast CHECK POINT # 2
  5. Transform into embedding Find Potential Customer Query Current Customer Vector

    Representation Query Vector (topK) Reply Customer List to Owner
  6. Vector Similarity Explained Euclidean distance Cosine similarity Dot product measures

    the straight-line distance between two points measures the angle between vectors measures the projection of one vector onto another
  7. Transform into embedding Results Inform Owner Embedding to Vector Query

    Vector Generate Image Narrowcast CHECK POINT # 3
  8. Transform into embedding Results Inform Owner Embedding to Vector Query

    Vector Generate Image Narrowcast CHECK POINT # 4
  9. Narrowcast
 to Customer Query Vector Embedding To Vector Push Message

    Empty Queue Query All Customer Store Vector Generate
 Image Reply Potential Customer List