Densifying Assumed-sparse Tensors: Improving Memory Efficiency and MPI Collective Performance during Tensor Accumulation for Parallelized Training of Neural Machine Translation Models.
Derya ÇavdarValeriu CodreanuCan KarakusJohn A. Lockman IIIDamian PodareanuVikram A. SaletoreAlexander SergeevDon D. Smith IIVictor SuthichaiQuy TaSrinivas VaradharajanLucas A. WilsonRengan XuPei YangPublished in: CoRR (2019)
Keyphrases
- machine translation
- high order
- cross lingual
- information extraction
- machine translation system
- natural language processing
- tensor factorization
- cross language information retrieval
- associative memory
- statistical machine translation
- diffusion tensor
- target language
- language independent
- finite state transducers
- language model
- tasks in natural language processing
- data mining
- evaluation metrics
- statistical models
- probabilistic model