Word Boundary Information Isn't Useful for Encoder Language Models.
Edward Gow-SmithDylan PhelpsHarish Tayyar MadabushiCarolina ScartonAline VillavicencioPublished in: CoRR (2024)
Keyphrases
- language model
- boundary information
- n gram
- translation model
- language modeling
- probabilistic model
- statistical language modeling
- image segmentation
- region growing
- probabilistic modeling
- information retrieval
- query expansion
- document retrieval
- speech recognition
- test collection
- retrieval model
- mixture model
- bit rate
- region merging
- word segmentation
- spoken term detection
- object segmentation
- bag of words
- motion estimation
- smoothing methods
- image regions
- relevance model
- spatial coherence
- language models for information retrieval
- expectation maximization
- image features
- information extraction
- object recognition