Matches in SemOpenAlex for { <https://semopenalex.org/work/W1506666725> ?p ?o ?g. }
Showing items 1 to 98 of
98
with 100 items per page.
- W1506666725 abstract "In HMM/DNN automatic speech recognition (ASR) systems, the DNNs model the posterior probabilities for triphone states. However, triphone states are unevenly distributed. In this situation, the training algorithm tends to converge to a local optimum more related to states with rich data than states with poor data. Thus, the imbalance of the training data decreases the ASR performances, especially for under-resourced languages. To deal with this issue, we explore a resampling technique, called “probabilistic sampling”, which can be seen as a linear smoothing between the original sampling and the uniform sampling. The effectiveness of the probabilistic sampling has been studied in two under-resourced ASR experiments. With the probabilistic sampling, the first experiment got a 6.3% relative phone error rate (PER) reduction compared to the conventional DNN baseline; the second experiment used shared-hidden-layer multilingual DNN as the baseline, and obtained a 4.9% relative PER reduction." @default.
- W1506666725 created "2016-06-24" @default.
- W1506666725 creator A5035445599 @default.
- W1506666725 creator A5047784934 @default.
- W1506666725 creator A5059968580 @default.
- W1506666725 creator A5066595006 @default.
- W1506666725 date "2015-07-01" @default.
- W1506666725 modified "2023-09-25" @default.
- W1506666725 title "Improving HMM/DNN in ASR of under-resourced languages using probabilistic sampling" @default.
- W1506666725 cites W1500890010 @default.
- W1506666725 cites W1554663460 @default.
- W1506666725 cites W1574715085 @default.
- W1506666725 cites W1978660892 @default.
- W1506666725 cites W1994606281 @default.
- W1506666725 cites W2025198378 @default.
- W1506666725 cites W2064244481 @default.
- W1506666725 cites W2091267617 @default.
- W1506666725 cites W2106440210 @default.
- W1506666725 cites W2107553890 @default.
- W1506666725 cites W2153682659 @default.
- W1506666725 cites W2160815625 @default.
- W1506666725 cites W2164499211 @default.
- W1506666725 cites W2394932179 @default.
- W1506666725 cites W97072897 @default.
- W1506666725 doi "https://doi.org/10.1109/chinasip.2015.7230354" @default.
- W1506666725 hasPublicationYear "2015" @default.
- W1506666725 type Work @default.
- W1506666725 sameAs 1506666725 @default.
- W1506666725 citedByCount "1" @default.
- W1506666725 countsByYear W15066667252017 @default.
- W1506666725 crossrefType "proceedings-article" @default.
- W1506666725 hasAuthorship W1506666725A5035445599 @default.
- W1506666725 hasAuthorship W1506666725A5047784934 @default.
- W1506666725 hasAuthorship W1506666725A5059968580 @default.
- W1506666725 hasAuthorship W1506666725A5066595006 @default.
- W1506666725 hasConcept C111335779 @default.
- W1506666725 hasConcept C138885662 @default.
- W1506666725 hasConcept C140779682 @default.
- W1506666725 hasConcept C150921843 @default.
- W1506666725 hasConcept C154945302 @default.
- W1506666725 hasConcept C23224414 @default.
- W1506666725 hasConcept C2524010 @default.
- W1506666725 hasConcept C2778707766 @default.
- W1506666725 hasConcept C28490314 @default.
- W1506666725 hasConcept C31972630 @default.
- W1506666725 hasConcept C33923547 @default.
- W1506666725 hasConcept C3770464 @default.
- W1506666725 hasConcept C40969351 @default.
- W1506666725 hasConcept C41008148 @default.
- W1506666725 hasConcept C41895202 @default.
- W1506666725 hasConcept C49937458 @default.
- W1506666725 hasConcept C76155785 @default.
- W1506666725 hasConcept C94915269 @default.
- W1506666725 hasConceptScore W1506666725C111335779 @default.
- W1506666725 hasConceptScore W1506666725C138885662 @default.
- W1506666725 hasConceptScore W1506666725C140779682 @default.
- W1506666725 hasConceptScore W1506666725C150921843 @default.
- W1506666725 hasConceptScore W1506666725C154945302 @default.
- W1506666725 hasConceptScore W1506666725C23224414 @default.
- W1506666725 hasConceptScore W1506666725C2524010 @default.
- W1506666725 hasConceptScore W1506666725C2778707766 @default.
- W1506666725 hasConceptScore W1506666725C28490314 @default.
- W1506666725 hasConceptScore W1506666725C31972630 @default.
- W1506666725 hasConceptScore W1506666725C33923547 @default.
- W1506666725 hasConceptScore W1506666725C3770464 @default.
- W1506666725 hasConceptScore W1506666725C40969351 @default.
- W1506666725 hasConceptScore W1506666725C41008148 @default.
- W1506666725 hasConceptScore W1506666725C41895202 @default.
- W1506666725 hasConceptScore W1506666725C49937458 @default.
- W1506666725 hasConceptScore W1506666725C76155785 @default.
- W1506666725 hasConceptScore W1506666725C94915269 @default.
- W1506666725 hasLocation W15066667251 @default.
- W1506666725 hasOpenAccess W1506666725 @default.
- W1506666725 hasPrimaryLocation W15066667251 @default.
- W1506666725 hasRelatedWork W1964863507 @default.
- W1506666725 hasRelatedWork W1986184096 @default.
- W1506666725 hasRelatedWork W2012257340 @default.
- W1506666725 hasRelatedWork W2016084804 @default.
- W1506666725 hasRelatedWork W2019973879 @default.
- W1506666725 hasRelatedWork W2053067333 @default.
- W1506666725 hasRelatedWork W2063689849 @default.
- W1506666725 hasRelatedWork W2066824132 @default.
- W1506666725 hasRelatedWork W2071310251 @default.
- W1506666725 hasRelatedWork W2075255969 @default.
- W1506666725 hasRelatedWork W2085377355 @default.
- W1506666725 hasRelatedWork W2162871885 @default.
- W1506666725 hasRelatedWork W2291439375 @default.
- W1506666725 hasRelatedWork W2291934070 @default.
- W1506666725 hasRelatedWork W2430252546 @default.
- W1506666725 hasRelatedWork W2605514952 @default.
- W1506666725 hasRelatedWork W2751522352 @default.
- W1506666725 hasRelatedWork W2767064007 @default.
- W1506666725 hasRelatedWork W2059610484 @default.
- W1506666725 hasRelatedWork W2157844668 @default.
- W1506666725 isParatext "false" @default.
- W1506666725 isRetracted "false" @default.
- W1506666725 magId "1506666725" @default.
- W1506666725 workType "article" @default.