Limited-Memory BFGS Optimization of Recurrent Neural Network Language Models for Speech Recognition.
Xunying LiuShansong LiuJinze ShaJianwei YuZhiyuan XuXie ChenHelen MengPublished in: ICASSP (2018)
Keyphrases
- limited memory
- speech recognition
- language model
- recurrent neural networks
- language modeling
- sliding window
- influence diagrams
- neural network
- document retrieval
- memory space
- probabilistic model
- n gram
- quasi newton method
- data streams
- real time
- automatic speech recognition
- retrieval model
- feed forward
- test collection
- speech signal
- query expansion
- speech recognition systems
- word error rate
- mixture model
- hidden markov models
- information retrieval
- handwriting recognition
- relevance model
- sensitivity analysis
- query terms
- quasi newton
- artificial neural networks
- document collections
- text categorization