I2MVFormer: Large Language Model Generated Multi-View Document Supervision for Zero-Shot Image Classification.
Muhammad Ferjad NaeemMuhammad Gul Zain Ali KhanYongqin XianMuhammad Zeshan AfzalDidier StrickerLuc Van GoolFederico TombariPublished in: CoRR (2022)
Keyphrases
- multi view
- language model
- image classification
- document retrieval
- ad hoc information retrieval
- document ranking
- information retrieval
- language modeling
- document length
- query terms
- single view
- vector space model
- word clouds
- n gram
- test collection
- probabilistic model
- query specific
- retrieval model
- d objects
- relevance model
- image representation
- language modeling framework
- query expansion
- bag of words
- three dimensional
- mixture model
- relevant documents
- retrieval systems
- information retrieval systems
- feature extraction
- co training
- semi supervised
- smoothing methods
- visual features
- translation model
- image features
- tf idf
- document clustering
- web documents
- topic modeling
- document collections
- keywords
- inter document similarities
- object categories
- computer vision