Structural Pruning of Pre-trained Language Models via Neural Architecture Search.
Aaron KleinJacek GolebiowskiXingchen MaValerio PerroneCédric ArchambeauPublished in: CoRR (2024)
Keyphrases
- language model
- neural architecture
- pre trained
- language modeling
- search space
- n gram
- probabilistic model
- retrieval model
- document retrieval
- speech recognition
- document ranking
- test collection
- information retrieval
- neural network
- query expansion
- training data
- feed forward
- relevance model
- similarity measure
- language models for information retrieval
- data sets
- smoothing methods
- machine learning
- learning algorithm
- activation function
- multi layer perceptron
- active learning
- data fusion
- state space
- support vector machine
- small number