DeepDiffusion: Unsupervised Learning of Retrieval-adapted Representations via Diffusion-based Ranking on Latent Feature Manifold.
Takahiko FuruyaRyutarou OhbuchiPublished in: CoRR (2021)
Keyphrases
- unsupervised learning
- manifold ranking
- information retrieval
- ranking algorithm
- vector representation
- feature ranking
- diffusion process
- information retrieval systems
- image database
- supervised learning
- manifold learning
- semi supervised
- feature vectors
- document ranking
- test collection
- latent variables
- ranking functions
- retrieval quality
- deep belief networks
- retrieval functions
- document retrieval
- retrieval systems
- image features
- heat diffusion
- image retrieval
- retrieval strategies
- learning to rank
- web search
- relevant documents
- relevance feedback
- euclidean space
- feature set
- active learning
- low dimensional
- machine learning
- high dimensional
- pairwise
- object recognition
- pattern recognition
- reinforcement learning