Joint Embedding Architecture
Definition
A neural network design where inputs are encoded to embeddings by encoder networks, and learning happens via comparison or prediction in embedding space rather than in the original input space (e.g., pixels). Avoids reconstruction but risks representation collapse without proper constraints.
Cast within
No topic-region cast yet — this appears once Joint Embedding Architecture's compiled claims are aligned into a topic region's argument tree.
Claims mentioning Joint Embedding Architecture (5)
Your notes
Loading notes...