The Unreasonable Effectiveness of Large Language-Vision Models for Source-free Video Domain Adaptation.
Giacomo ZaraAlessandro ContiSubhankar RoyStéphane LathuilièrePaolo RotaElisa RicciPublished in: CoRR (2023)
Keyphrases
- domain adaptation
- probabilistic model
- multiple sources
- computer vision
- target domain
- model selection
- natural language
- video sequences
- error rate
- transfer learning
- database systems
- e learning
- covariate shift
- semi supervised
- general purpose
- multi view
- training data
- labeled data
- semi supervised learning
- co training
- databases
- target language