The Embeddings That Came in From the Cold: Improving Vectors for New and Rare Products with Content-Based Inference.
Jacopo TagliabueBingqing YuFederico BianchiPublished in: RecSys (2020)
Keyphrases
- vector space
- image retrieval
- multimedia
- bayesian networks
- feature vectors
- euclidean space
- low dimensional
- bayesian inference
- hilbert space
- inference process
- probabilistic inference
- belief networks
- metadata
- inference engine
- dynamic bayesian networks
- image indexing
- production line
- input pattern
- rare events
- distance measure
- product line
- customer requirements