Re-ViLM: Retrieval-Augmented Visual Language Model for Zero and Few-Shot Image Captioning.
Zhuolin YangWei PingZihan LiuVijay KorthikantiWeili NieDe-An HuangLinxi FanZhiding YuShiyi LanBo LiMohammad ShoeybiMing-Yu LiuYuke ZhuBryan CatanzaroChaowei XiaoAnima AnandkumarPublished in: EMNLP (Findings) (2023)
Keyphrases
- language model
- image retrieval
- document retrieval
- retrieval model
- information retrieval
- language modeling
- query expansion
- visual features
- test collection
- low level
- image features
- image classification
- query terms
- n gram
- image content
- ad hoc information retrieval
- image representation
- cross language retrieval
- smoothing methods
- document ranking
- language models for information retrieval
- image segmentation
- speech recognition
- probabilistic model
- statistical language models
- query specific
- mixture model
- information retrieval systems
- visual information
- context sensitive
- image collections
- retrieval method
- language modelling
- image database
- news video
- relevance model
- vector space model
- document length
- statistical language modeling
- bayesian framework