Matches in SemOpenAlex for { <https://semopenalex.org/work/W2649558613> ?p ?o ?g. }
- W2649558613 abstract "Leveraging upon transfer learning, we distill the knowledge in a conventional wide and deep neural network (DNN) into a narrower yet deeper model with fewer parameters and comparable system performance for speech enhancement. We present three transfer-learning solutions to accomplish our goal. First, the knowledge embedded in the form of the output values of a high-performance DNN is used to guide the training of a smaller DNN model in sequential transfer learning. In the second multi-task transfer learning solution, the smaller DNN is trained to learn the output value of the larger DNN, and the speech enhancement task in parallel. Finally, a progressive stacking transfer learning is accomplished through multi-task learning, and DNN stacking. Our experimental evidences demonstrate 5 times parameter reduction while maintaining similar enhancement performance with the proposed framework." @default.
- W2649558613 created "2017-06-30" @default.
- W2649558613 creator A5011008665 @default.
- W2649558613 creator A5012954193 @default.
- W2649558613 creator A5066868860 @default.
- W2649558613 creator A5079659476 @default.
- W2649558613 creator A5083304846 @default.
- W2649558613 date "2017-03-01" @default.
- W2649558613 modified "2023-09-26" @default.
- W2649558613 title "A transfer learning and progressive stacking approach to reducing deep model sizes with an application to speech enhancement" @default.
- W2649558613 cites W1533861849 @default.
- W2649558613 cites W1686810756 @default.
- W2649558613 cites W1904365287 @default.
- W2649558613 cites W1974387177 @default.
- W2649558613 cites W2044893557 @default.
- W2649558613 cites W2078528584 @default.
- W2649558613 cites W2100969003 @default.
- W2649558613 cites W2114766824 @default.
- W2649558613 cites W2130306094 @default.
- W2649558613 cites W2139008940 @default.
- W2649558613 cites W2143612262 @default.
- W2649558613 cites W2147768505 @default.
- W2649558613 cites W2156150815 @default.
- W2649558613 cites W2165698076 @default.
- W2649558613 cites W2294370754 @default.
- W2649558613 cites W2296516251 @default.
- W2649558613 cites W2304609584 @default.
- W2649558613 cites W2402040300 @default.
- W2649558613 cites W2407793339 @default.
- W2649558613 cites W2520176975 @default.
- W2649558613 cites W2952432176 @default.
- W2649558613 cites W2962988160 @default.
- W2649558613 cites W2963674932 @default.
- W2649558613 cites W2964299589 @default.
- W2649558613 cites W1614862348 @default.
- W2649558613 doi "https://doi.org/10.1109/icassp.2017.7953223" @default.
- W2649558613 hasPublicationYear "2017" @default.
- W2649558613 type Work @default.
- W2649558613 sameAs 2649558613 @default.
- W2649558613 citedByCount "12" @default.
- W2649558613 countsByYear W26495586132017 @default.
- W2649558613 countsByYear W26495586132018 @default.
- W2649558613 countsByYear W26495586132019 @default.
- W2649558613 countsByYear W26495586132020 @default.
- W2649558613 countsByYear W26495586132021 @default.
- W2649558613 crossrefType "proceedings-article" @default.
- W2649558613 hasAuthorship W2649558613A5011008665 @default.
- W2649558613 hasAuthorship W2649558613A5012954193 @default.
- W2649558613 hasAuthorship W2649558613A5066868860 @default.
- W2649558613 hasAuthorship W2649558613A5079659476 @default.
- W2649558613 hasAuthorship W2649558613A5083304846 @default.
- W2649558613 hasConcept C108583219 @default.
- W2649558613 hasConcept C111335779 @default.
- W2649558613 hasConcept C119857082 @default.
- W2649558613 hasConcept C121332964 @default.
- W2649558613 hasConcept C127413603 @default.
- W2649558613 hasConcept C150899416 @default.
- W2649558613 hasConcept C154945302 @default.
- W2649558613 hasConcept C163294075 @default.
- W2649558613 hasConcept C173608175 @default.
- W2649558613 hasConcept C201995342 @default.
- W2649558613 hasConcept C2524010 @default.
- W2649558613 hasConcept C2776175482 @default.
- W2649558613 hasConcept C2776182073 @default.
- W2649558613 hasConcept C2776960227 @default.
- W2649558613 hasConcept C2780451532 @default.
- W2649558613 hasConcept C28006648 @default.
- W2649558613 hasConcept C28490314 @default.
- W2649558613 hasConcept C33347731 @default.
- W2649558613 hasConcept C33923547 @default.
- W2649558613 hasConcept C41008148 @default.
- W2649558613 hasConcept C46141821 @default.
- W2649558613 hasConcept C50644808 @default.
- W2649558613 hasConcept C56739046 @default.
- W2649558613 hasConceptScore W2649558613C108583219 @default.
- W2649558613 hasConceptScore W2649558613C111335779 @default.
- W2649558613 hasConceptScore W2649558613C119857082 @default.
- W2649558613 hasConceptScore W2649558613C121332964 @default.
- W2649558613 hasConceptScore W2649558613C127413603 @default.
- W2649558613 hasConceptScore W2649558613C150899416 @default.
- W2649558613 hasConceptScore W2649558613C154945302 @default.
- W2649558613 hasConceptScore W2649558613C163294075 @default.
- W2649558613 hasConceptScore W2649558613C173608175 @default.
- W2649558613 hasConceptScore W2649558613C201995342 @default.
- W2649558613 hasConceptScore W2649558613C2524010 @default.
- W2649558613 hasConceptScore W2649558613C2776175482 @default.
- W2649558613 hasConceptScore W2649558613C2776182073 @default.
- W2649558613 hasConceptScore W2649558613C2776960227 @default.
- W2649558613 hasConceptScore W2649558613C2780451532 @default.
- W2649558613 hasConceptScore W2649558613C28006648 @default.
- W2649558613 hasConceptScore W2649558613C28490314 @default.
- W2649558613 hasConceptScore W2649558613C33347731 @default.
- W2649558613 hasConceptScore W2649558613C33923547 @default.
- W2649558613 hasConceptScore W2649558613C41008148 @default.
- W2649558613 hasConceptScore W2649558613C46141821 @default.
- W2649558613 hasConceptScore W2649558613C50644808 @default.
- W2649558613 hasConceptScore W2649558613C56739046 @default.
- W2649558613 hasLocation W26495586131 @default.
- W2649558613 hasOpenAccess W2649558613 @default.
- W2649558613 hasPrimaryLocation W26495586131 @default.