Weakly-supervised Audio Separation via Bi-modal Semantic Similarity.
Tanvir MahmudSaeed AmizadehKazuhito KoishidaDiana MarculescuPublished in: ICLR (2024)
Keyphrases
- semantic similarity
- weakly supervised
- co occurrence
- relation extraction
- semantic features
- topic models
- object class
- named entities
- wordnet
- semantic information
- superpixels
- similarity measure
- semi supervised
- visual information
- object detection
- object recognition
- word sense disambiguation
- databases
- ground truth
- multiscale
- visual similarity
- multi modal
- natural language processing
- information extraction
- knowledge representation
- expert systems
- metadata
- machine learning