EDGE OSDEFINITION
A joint embedding is an example of machine learning where you can embed 2 or more data types, or inputs, into a vector space. This allows information about the relationships between the data types to be examined. A model could be trained to embed both images and text to describe them in an embedding space. This would mean that semantically similar items would be closer in the embedding space. It is useful in multi-modal AI to allow systems to link different data types, without needing exact keywords.
Importance
- Cross Modal Understanding: Allows information to be linked between different data types (text, images).
- Semantic Similarity: Can identify similar content using vectors.
- Information Retrieval: Search over different modalities.
- AI Integration: Can be used within multi-modal applications.
- Content Matching: Can match descriptions, documents and media with similar information.
Use Cases
- Search for images based on natural language queries.
- Match product images with descriptions from a catalog.
- Search for relevant diagrams/screenshots from a technical document.
- Build a multi-modal search system for your enterprise knowledge base.
- Link text queries with video frames/audio within supported AI systems.
See it running on your own hardware
Vanik Edge OS runs AI models, agents and apps inside your building, so your data stays there.