Quantifying the Visual Concreteness of Words and Topics in Multimodal Datasets.
Jack HesselDavid M. MimnoLillian LeePublished in: NAACL-HLT (2018)
Keyphrases
- keywords
- text documents
- multimodal information
- visual features
- n gram
- latent topics
- topic hierarchy
- topic models
- cross modal
- database
- multi modal
- information retrieval
- visual information
- auto annotation
- related topics
- statistical topic models
- latent dirichlet allocation
- related words
- text corpora
- visual data
- data sets
- word recognition
- multiword
- topic detection
- semantic information
- information extraction
- hidden markov models
- multimedia