Matches in SemOpenAlex for { <https://semopenalex.org/work/W3081034749> ?p ?o ?g. }
Showing items 1 to 76 of
76
with 100 items per page.
- W3081034749 abstract "Deep neural networks (DNNs) have been proved to be popular and successful in our daily life. For mobile devices with the services from DNN, it is needed to equip a DNN accelerator to employ fast and accurate DNN models. Different DNN accelerators may have different accuracy performance, which directly affects the user experience. To measure the accuracy of the DNN accelerators, there only exists indirect approach by comparing to the third-party platform with 32-bit float format. In this paper, metamorphic testing is employed to measure the accuracy of the DNN accelerator for the first time. Specifically, our scheme utilizes the mathematical properties of the operators in deep neural networks to measure the accuracy of the DNN accelerator, getting rid of the dependence on the third-party platform. Besides, we design an objective metric to quantitatively measure the accuracy performance. It can be used not only to facilitate the direct competition between different accelerators but also to detect defects in the implementation of the DNN accelerator. An experiment on two well-known DNN accelerators, HiAI and Snapdragon Neural Processing Engine (SNPE), is conducted, and the results show that the metamorphic distance of the former is smaller than that of the latter in terms of their accuracy performance with 16-bit float format on typical Ops, such as softmax and convolution." @default.
- W3081034749 created "2020-09-01" @default.
- W3081034749 creator A5010026045 @default.
- W3081034749 creator A5014629413 @default.
- W3081034749 creator A5021452379 @default.
- W3081034749 creator A5048567667 @default.
- W3081034749 creator A5075174750 @default.
- W3081034749 creator A5077629501 @default.
- W3081034749 date "2020-08-01" @default.
- W3081034749 modified "2023-10-16" @default.
- W3081034749 title "Accuracy Measurement of Deep Neural Network Accelerator via Metamorphic Testing" @default.
- W3081034749 cites W2049434052 @default.
- W3081034749 cites W2110441383 @default.
- W3081034749 cites W2155893237 @default.
- W3081034749 cites W2194775991 @default.
- W3081034749 cites W2324595780 @default.
- W3081034749 cites W2782311202 @default.
- W3081034749 cites W2789057423 @default.
- W3081034749 cites W2939371957 @default.
- W3081034749 cites W2963037989 @default.
- W3081034749 cites W2967997213 @default.
- W3081034749 doi "https://doi.org/10.1109/aitest49225.2020.00015" @default.
- W3081034749 hasPublicationYear "2020" @default.
- W3081034749 type Work @default.
- W3081034749 sameAs 3081034749 @default.
- W3081034749 citedByCount "2" @default.
- W3081034749 countsByYear W30810347492021 @default.
- W3081034749 countsByYear W30810347492023 @default.
- W3081034749 crossrefType "proceedings-article" @default.
- W3081034749 hasAuthorship W3081034749A5010026045 @default.
- W3081034749 hasAuthorship W3081034749A5014629413 @default.
- W3081034749 hasAuthorship W3081034749A5021452379 @default.
- W3081034749 hasAuthorship W3081034749A5048567667 @default.
- W3081034749 hasAuthorship W3081034749A5075174750 @default.
- W3081034749 hasAuthorship W3081034749A5077629501 @default.
- W3081034749 hasConcept C108583219 @default.
- W3081034749 hasConcept C124101348 @default.
- W3081034749 hasConcept C154945302 @default.
- W3081034749 hasConcept C162324750 @default.
- W3081034749 hasConcept C176217482 @default.
- W3081034749 hasConcept C188441871 @default.
- W3081034749 hasConcept C21547014 @default.
- W3081034749 hasConcept C2780009758 @default.
- W3081034749 hasConcept C41008148 @default.
- W3081034749 hasConcept C45347329 @default.
- W3081034749 hasConcept C50644808 @default.
- W3081034749 hasConcept C81363708 @default.
- W3081034749 hasConceptScore W3081034749C108583219 @default.
- W3081034749 hasConceptScore W3081034749C124101348 @default.
- W3081034749 hasConceptScore W3081034749C154945302 @default.
- W3081034749 hasConceptScore W3081034749C162324750 @default.
- W3081034749 hasConceptScore W3081034749C176217482 @default.
- W3081034749 hasConceptScore W3081034749C188441871 @default.
- W3081034749 hasConceptScore W3081034749C21547014 @default.
- W3081034749 hasConceptScore W3081034749C2780009758 @default.
- W3081034749 hasConceptScore W3081034749C41008148 @default.
- W3081034749 hasConceptScore W3081034749C45347329 @default.
- W3081034749 hasConceptScore W3081034749C50644808 @default.
- W3081034749 hasConceptScore W3081034749C81363708 @default.
- W3081034749 hasLocation W30810347491 @default.
- W3081034749 hasOpenAccess W3081034749 @default.
- W3081034749 hasPrimaryLocation W30810347491 @default.
- W3081034749 hasRelatedWork W2731899572 @default.
- W3081034749 hasRelatedWork W2913997398 @default.
- W3081034749 hasRelatedWork W2977314777 @default.
- W3081034749 hasRelatedWork W2999805992 @default.
- W3081034749 hasRelatedWork W3116150086 @default.
- W3081034749 hasRelatedWork W3133861977 @default.
- W3081034749 hasRelatedWork W4200173597 @default.
- W3081034749 hasRelatedWork W4291897433 @default.
- W3081034749 hasRelatedWork W4312417841 @default.
- W3081034749 hasRelatedWork W4321369474 @default.
- W3081034749 isParatext "false" @default.
- W3081034749 isRetracted "false" @default.
- W3081034749 magId "3081034749" @default.
- W3081034749 workType "article" @default.