FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287322404> ?p ?o ?g. }

• W4287322404 abstract "Soft error, namely silent corruption of signal or datum in a computer system, cannot be caverlierly ignored as compute and communication density grow exponentially. Soft error detection has been studied in the context of enterprise computing, high-performance computing and more recently in convolutional neural networks related to autonomous driving.Deep learning recommendation systems (DLRMs) have by now become ubiquitous and serve billions of users per day. Nevertheless, DLRM-specific soft error detection methods are hitherto missing. To fill the gap, this paper presents the first set of soft-error detection methods for low-precision quantized-arithmetic operators in DLRM including general matrix multiplication (GEMM) and EmbeddingBag. A practical method must detect error and do so with low overhead lest reduced inference speed degrades user experience. Exploiting the characteristics of both quantized arithmetic and the operators, we achieved more than 95% detection accuracy for GEMM with an overhead below 20%. For EmbeddingBag, we achieved 99% effectiveness in significant-bit-flips detection with less than 10% of false positives, while keeping overhead below 26%." @default.
• W4287322404 created "2022-07-25" @default.
• W4287322404 creator A5003198990 @default.
• W4287322404 creator A5018336899 @default.
• W4287322404 creator A5046705985 @default.
• W4287322404 creator A5051151162 @default.
• W4287322404 creator A5060661843 @default.
• W4287322404 creator A5061737717 @default.
• W4287322404 creator A5069883905 @default.
• W4287322404 date "2022-12-17" @default.
• W4287322404 modified "2023-10-01" @default.
• W4287322404 title "Efficient Soft-Error Detection for Low-precision Deep Learning Recommendation Models" @default.
• W4287322404 cites W2034593585 @default.
• W4287322404 cites W2035492130 @default.
• W4287322404 cites W2052455844 @default.
• W4287322404 cites W2083613288 @default.
• W4287322404 cites W2108434125 @default.
• W4287322404 cites W2152212524 @default.
• W4287322404 cites W2152652532 @default.
• W4287322404 cites W2156514327 @default.
• W4287322404 cites W2296204683 @default.
• W4287322404 cites W2412349256 @default.
• W4287322404 cites W2418331349 @default.
• W4287322404 cites W2485331474 @default.
• W4287322404 cites W2580243656 @default.
• W4287322404 cites W2767260595 @default.
• W4287322404 cites W2767694495 @default.
• W4287322404 cites W2770542984 @default.
• W4287322404 cites W2901848761 @default.
• W4287322404 cites W2963122961 @default.
• W4287322404 cites W2984020950 @default.
• W4287322404 cites W3007788310 @default.
• W4287322404 cites W3105862567 @default.
• W4287322404 cites W3130104841 @default.
• W4287322404 cites W4230550535 @default.
• W4287322404 cites W4248621761 @default.
• W4287322404 cites W4233447799 @default.
• W4287322404 doi "https://doi.org/10.1109/bigdata55660.2022.10020972" @default.
• W4287322404 hasPublicationYear "2022" @default.
• W4287322404 type Work @default.
• W4287322404 citedByCount "0" @default.
• W4287322404 crossrefType "proceedings-article" @default.
• W4287322404 hasAuthorship W4287322404A5003198990 @default.
• W4287322404 hasAuthorship W4287322404A5018336899 @default.
• W4287322404 hasAuthorship W4287322404A5046705985 @default.
• W4287322404 hasAuthorship W4287322404A5051151162 @default.
• W4287322404 hasAuthorship W4287322404A5060661843 @default.
• W4287322404 hasAuthorship W4287322404A5061737717 @default.
• W4287322404 hasAuthorship W4287322404A5069883905 @default.
• W4287322404 hasBestOaLocation W42873224042 @default.
• W4287322404 hasConcept C103088060 @default.
• W4287322404 hasConcept C108583219 @default.
• W4287322404 hasConcept C111919701 @default.
• W4287322404 hasConcept C113775141 @default.
• W4287322404 hasConcept C11413529 @default.
• W4287322404 hasConcept C121332964 @default.
• W4287322404 hasConcept C151730666 @default.
• W4287322404 hasConcept C154945302 @default.
• W4287322404 hasConcept C17349429 @default.
• W4287322404 hasConcept C2776214188 @default.
• W4287322404 hasConcept C2779343474 @default.
• W4287322404 hasConcept C2779960059 @default.
• W4287322404 hasConcept C41008148 @default.
• W4287322404 hasConcept C62520636 @default.
• W4287322404 hasConcept C64869954 @default.
• W4287322404 hasConcept C81363708 @default.
• W4287322404 hasConcept C84114770 @default.
• W4287322404 hasConcept C86803240 @default.
• W4287322404 hasConceptScore W4287322404C103088060 @default.
• W4287322404 hasConceptScore W4287322404C108583219 @default.
• W4287322404 hasConceptScore W4287322404C111919701 @default.
• W4287322404 hasConceptScore W4287322404C113775141 @default.
• W4287322404 hasConceptScore W4287322404C11413529 @default.
• W4287322404 hasConceptScore W4287322404C121332964 @default.
• W4287322404 hasConceptScore W4287322404C151730666 @default.
• W4287322404 hasConceptScore W4287322404C154945302 @default.
• W4287322404 hasConceptScore W4287322404C17349429 @default.
• W4287322404 hasConceptScore W4287322404C2776214188 @default.
• W4287322404 hasConceptScore W4287322404C2779343474 @default.
• W4287322404 hasConceptScore W4287322404C2779960059 @default.
• W4287322404 hasConceptScore W4287322404C41008148 @default.
• W4287322404 hasConceptScore W4287322404C62520636 @default.
• W4287322404 hasConceptScore W4287322404C64869954 @default.
• W4287322404 hasConceptScore W4287322404C81363708 @default.
• W4287322404 hasConceptScore W4287322404C84114770 @default.
• W4287322404 hasConceptScore W4287322404C86803240 @default.
• W4287322404 hasFunder F4320306084 @default.
• W4287322404 hasFunder F4320332359 @default.
• W4287322404 hasFunder F4320337506 @default.
• W4287322404 hasLocation W42873224041 @default.
• W4287322404 hasLocation W42873224042 @default.
• W4287322404 hasOpenAccess W4287322404 @default.
• W4287322404 hasPrimaryLocation W42873224041 @default.
• W4287322404 hasRelatedWork W1585772192 @default.
• W4287322404 hasRelatedWork W2389317155 @default.
• W4287322404 hasRelatedWork W2731899572 @default.
• W4287322404 hasRelatedWork W2973942527 @default.
• W4287322404 hasRelatedWork W3133174490 @default.
• W4287322404 hasRelatedWork W3133861977 @default.
• W4287322404 hasRelatedWork W3134064223 @default.

• next