FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4311263200> ?p ?o ?g. }

• W4311263200 abstract "Deep neural net works have brought significant innovations in many domains, such as computer vision, natural language processing, speech recognition, and intelligent decision. To this end, many new operators have also been developed to attain better accuracy in the specific domain, such as the operators about anchors in object detection and operators about agent-environment interaction in reinforcement learning. In this paper, we identify that the new domain-specific operators, which have no corresponding implementation in the device compute library (such as cuDNN, and PyTorch ATen) and have to resort to the Python interpreter, will introduce large amounts of challenges in deep learning training and deployment. We name these operators long-tail operators, inspired by the meaning of the long-tail phenomenon in business and statistics.As such, researchers have developed kinds of deep learning compilers, such as XLA, TorchScript, $mathcal{J}{A X}$, and TVM, to solve the challenges incurred by long-tail operators. Since there are a lot of complex syntax features in the long-tail operators, providing a well-designed benchmark suite to assess and profile the deep learning compilers is of much importance. Unfortunately, there have been no representative benchmark suites that can take a solid understanding and analysis of long-tail operators. In this paper, we propose LongTail-Bench <sup xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>1</sup> , a benchmark suite for domain-specific operators in deep learning. To help researchers study deep learning compilers, LongTail-Bench collects more than 100 representative operators. We first perform a comprehensive analysis and show that LongTail-Bench can cover a wide range of syntax features. Then, we conduct a throughout evaluation of the mainstream deep learning compilers from multiple aspects, including training performance, deploy performance, comparison with handwritten code, and performance on the emerging deep learning accelerator. Finally, we take a research example to demonstrate that our benchmark suite can help to support research. <sup xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>1</sup> Corresponding Author" @default.
• W4311263200 created "2022-12-25" @default.
• W4311263200 creator A5010087030 @default.
• W4311263200 creator A5010817202 @default.
• W4311263200 creator A5031706167 @default.
• W4311263200 creator A5049910854 @default.
• W4311263200 creator A5053716293 @default.
• W4311263200 creator A5062964669 @default.
• W4311263200 creator A5085852932 @default.
• W4311263200 date "2022-11-01" @default.
• W4311263200 modified "2023-09-25" @default.
• W4311263200 title "LongTail-Bench: A Benchmark Suite for Domain-Specific Operators in Deep Learning" @default.
• W4311263200 cites W1536680647 @default.
• W4311263200 cites W1982205631 @default.
• W4311263200 cites W2002555321 @default.
• W4311263200 cites W2010452422 @default.
• W4311263200 cites W2117130368 @default.
• W4311263200 cites W2123442489 @default.
• W4311263200 cites W2160815625 @default.
• W4311263200 cites W2194775991 @default.
• W4311263200 cites W2565639579 @default.
• W4311263200 cites W2593649365 @default.
• W4311263200 cites W2883457665 @default.
• W4311263200 cites W2900082550 @default.
• W4311263200 cites W2911770499 @default.
• W4311263200 cites W2913790721 @default.
• W4311263200 cites W2962766617 @default.
• W4311263200 cites W2963150697 @default.
• W4311263200 cites W2963299996 @default.
• W4311263200 cites W2963351448 @default.
• W4311263200 cites W2963402592 @default.
• W4311263200 cites W2963674387 @default.
• W4311263200 cites W2964330541 @default.
• W4311263200 cites W2997747012 @default.
• W4311263200 cites W2998506323 @default.
• W4311263200 cites W3008378296 @default.
• W4311263200 cites W3098379356 @default.
• W4311263200 cites W3126815129 @default.
• W4311263200 cites W4206792915 @default.
• W4311263200 cites W4251637954 @default.
• W4311263200 doi "https://doi.org/10.1109/iiswc55918.2022.00032" @default.
• W4311263200 hasPublicationYear "2022" @default.
• W4311263200 type Work @default.
• W4311263200 citedByCount "0" @default.
• W4311263200 crossrefType "proceedings-article" @default.
• W4311263200 hasAuthorship W4311263200A5010087030 @default.
• W4311263200 hasAuthorship W4311263200A5010817202 @default.
• W4311263200 hasAuthorship W4311263200A5031706167 @default.
• W4311263200 hasAuthorship W4311263200A5049910854 @default.
• W4311263200 hasAuthorship W4311263200A5053716293 @default.
• W4311263200 hasAuthorship W4311263200A5062964669 @default.
• W4311263200 hasAuthorship W4311263200A5085852932 @default.
• W4311263200 hasConcept C108583219 @default.
• W4311263200 hasConcept C119857082 @default.
• W4311263200 hasConcept C13280743 @default.
• W4311263200 hasConcept C134306372 @default.
• W4311263200 hasConcept C154945302 @default.
• W4311263200 hasConcept C166957645 @default.
• W4311263200 hasConcept C169590947 @default.
• W4311263200 hasConcept C185798385 @default.
• W4311263200 hasConcept C199360897 @default.
• W4311263200 hasConcept C205649164 @default.
• W4311263200 hasConcept C33923547 @default.
• W4311263200 hasConcept C36503486 @default.
• W4311263200 hasConcept C41008148 @default.
• W4311263200 hasConcept C519991488 @default.
• W4311263200 hasConcept C79581498 @default.
• W4311263200 hasConcept C95457728 @default.
• W4311263200 hasConceptScore W4311263200C108583219 @default.
• W4311263200 hasConceptScore W4311263200C119857082 @default.
• W4311263200 hasConceptScore W4311263200C13280743 @default.
• W4311263200 hasConceptScore W4311263200C134306372 @default.
• W4311263200 hasConceptScore W4311263200C154945302 @default.
• W4311263200 hasConceptScore W4311263200C166957645 @default.
• W4311263200 hasConceptScore W4311263200C169590947 @default.
• W4311263200 hasConceptScore W4311263200C185798385 @default.
• W4311263200 hasConceptScore W4311263200C199360897 @default.
• W4311263200 hasConceptScore W4311263200C205649164 @default.
• W4311263200 hasConceptScore W4311263200C33923547 @default.
• W4311263200 hasConceptScore W4311263200C36503486 @default.
• W4311263200 hasConceptScore W4311263200C41008148 @default.
• W4311263200 hasConceptScore W4311263200C519991488 @default.
• W4311263200 hasConceptScore W4311263200C79581498 @default.
• W4311263200 hasConceptScore W4311263200C95457728 @default.
• W4311263200 hasLocation W43112632001 @default.
• W4311263200 hasOpenAccess W4311263200 @default.
• W4311263200 hasPrimaryLocation W43112632001 @default.
• W4311263200 hasRelatedWork W1497385637 @default.
• W4311263200 hasRelatedWork W1527862632 @default.
• W4311263200 hasRelatedWork W1608659702 @default.
• W4311263200 hasRelatedWork W1977285665 @default.
• W4311263200 hasRelatedWork W2007728761 @default.
• W4311263200 hasRelatedWork W2529681551 @default.
• W4311263200 hasRelatedWork W2740990710 @default.
• W4311263200 hasRelatedWork W3017187763 @default.
• W4311263200 hasRelatedWork W3087956025 @default.
• W4311263200 hasRelatedWork W4321365483 @default.
• W4311263200 isParatext "false" @default.
• W4311263200 isRetracted "false" @default.
• W4311263200 workType "article" @default.