eDKM: An Efficient and Accurate Train-Time Weight Clustering for Large Language Models.
Minsik ChoKeivan Alizadeh-VahidQichen FuSaurabh AdyaCarlo C. del MundoMohammad RastegariDevang NaikPeter ZatloukalPublished in: IEEE Comput. Archit. Lett. (2024)
Keyphrases
- language model
- language modeling
- n gram
- probabilistic model
- speech recognition
- retrieval model
- document retrieval
- query expansion
- information retrieval
- clustering algorithm
- statistical language models
- test collection
- language modelling
- smoothing methods
- clustering method
- relevance model
- context sensitive
- hierarchical clustering
- ad hoc information retrieval
- document ranking
- spectral clustering
- data clustering
- document clustering
- vector space model
- query terms
- mixture model
- k means
- text retrieval
- pseudo relevance feedback
- machine learning
- cross lingual
- weighting scheme
- document level
- co occurrence
- feature selection
- okapi bm