ROME: Evaluating Pre-trained Vision-Language Models on Reasoning beyond Visual Common Sense.
Kankan ZhouEason LaiWei Bin Au YeongKyriakos MouratidisJing JiangPublished in: EMNLP (Findings) (2023)
Keyphrases
- language model
- pre trained
- language modeling
- probabilistic model
- speech recognition
- document retrieval
- query expansion
- n gram
- language modelling
- retrieval model
- information retrieval
- visual features
- training data
- real time
- statistical language models
- test collection
- visual information
- relevance model
- document ranking
- computer vision
- training examples
- em algorithm
- semi supervised
- prior knowledge
- video sequences
- language models for information retrieval
- spoken term detection