MAGNET: Improving the Multilingual Fairness of Language Models with Adaptive Gradient-Based Tokenization.
Orevaoghene AhiaSachin KumarHila GonenValentin HoffmanTomasz LimisiewiczYulia TsvetkovNoah A. SmithPublished in: CoRR (2024)
Keyphrases
- language model
- language modeling
- n gram
- probabilistic model
- retrieval model
- cross lingual
- language independent
- information retrieval
- language modelling
- speech recognition
- document retrieval
- query expansion
- ad hoc information retrieval
- test collection
- statistical language models
- digital libraries
- pseudo relevance feedback
- language models for information retrieval
- vector space model
- context sensitive
- document ranking
- smoothing methods
- named entities
- relevance model
- document length
- feature selection
- machine learning
- query terms
- translation model
- term dependencies
- information retrieval systems