Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287328441> ?p ?o ?g. }
Showing items 1 to 75 of
75
with 100 items per page.
- W4287328441 abstract "Language models (LMs) for virtual assistants (VAs) are typically trained on large amounts of data, resulting in prohibitively large models which require excessive memory and/or cannot be used to serve user requests in real-time. Entropy pruning results in smaller models but with significant degradation of effectiveness in the tail of the user request distribution. We customize entropy pruning by allowing for a keep list of infrequent n-grams that require a more relaxed pruning threshold, and propose three methods to construct the keep list. Each method has its own advantages and disadvantages with respect to LM size, ASR accuracy and cost of constructing the keep list. Our best LM gives 8% average Word Error Rate (WER) reduction on a targeted test set, but is 3 times larger than the baseline. We also propose discriminative methods to reduce the size of the LM while retaining the majority of the WER gains achieved by the largest LM." @default.
- W4287328441 created "2022-07-25" @default.
- W4287328441 creator A5029549568 @default.
- W4287328441 creator A5048094198 @default.
- W4287328441 creator A5060527820 @default.
- W4287328441 creator A5084609544 @default.
- W4287328441 creator A5091396390 @default.
- W4287328441 date "2021-02-14" @default.
- W4287328441 modified "2023-09-24" @default.
- W4287328441 title "Error-driven Pruning of Language Models for Virtual Assistants" @default.
- W4287328441 doi "https://doi.org/10.48550/arxiv.2102.07219" @default.
- W4287328441 hasPublicationYear "2021" @default.
- W4287328441 type Work @default.
- W4287328441 citedByCount "0" @default.
- W4287328441 crossrefType "posted-content" @default.
- W4287328441 hasAuthorship W4287328441A5029549568 @default.
- W4287328441 hasAuthorship W4287328441A5048094198 @default.
- W4287328441 hasAuthorship W4287328441A5060527820 @default.
- W4287328441 hasAuthorship W4287328441A5084609544 @default.
- W4287328441 hasAuthorship W4287328441A5091396390 @default.
- W4287328441 hasBestOaLocation W42873284411 @default.
- W4287328441 hasConcept C106301342 @default.
- W4287328441 hasConcept C108010975 @default.
- W4287328441 hasConcept C111335779 @default.
- W4287328441 hasConcept C119857082 @default.
- W4287328441 hasConcept C121332964 @default.
- W4287328441 hasConcept C137293760 @default.
- W4287328441 hasConcept C153180895 @default.
- W4287328441 hasConcept C154945302 @default.
- W4287328441 hasConcept C167981619 @default.
- W4287328441 hasConcept C2524010 @default.
- W4287328441 hasConcept C28490314 @default.
- W4287328441 hasConcept C33923547 @default.
- W4287328441 hasConcept C40969351 @default.
- W4287328441 hasConcept C41008148 @default.
- W4287328441 hasConcept C62520636 @default.
- W4287328441 hasConcept C6557445 @default.
- W4287328441 hasConcept C86803240 @default.
- W4287328441 hasConcept C90805587 @default.
- W4287328441 hasConcept C97931131 @default.
- W4287328441 hasConceptScore W4287328441C106301342 @default.
- W4287328441 hasConceptScore W4287328441C108010975 @default.
- W4287328441 hasConceptScore W4287328441C111335779 @default.
- W4287328441 hasConceptScore W4287328441C119857082 @default.
- W4287328441 hasConceptScore W4287328441C121332964 @default.
- W4287328441 hasConceptScore W4287328441C137293760 @default.
- W4287328441 hasConceptScore W4287328441C153180895 @default.
- W4287328441 hasConceptScore W4287328441C154945302 @default.
- W4287328441 hasConceptScore W4287328441C167981619 @default.
- W4287328441 hasConceptScore W4287328441C2524010 @default.
- W4287328441 hasConceptScore W4287328441C28490314 @default.
- W4287328441 hasConceptScore W4287328441C33923547 @default.
- W4287328441 hasConceptScore W4287328441C40969351 @default.
- W4287328441 hasConceptScore W4287328441C41008148 @default.
- W4287328441 hasConceptScore W4287328441C62520636 @default.
- W4287328441 hasConceptScore W4287328441C6557445 @default.
- W4287328441 hasConceptScore W4287328441C86803240 @default.
- W4287328441 hasConceptScore W4287328441C90805587 @default.
- W4287328441 hasConceptScore W4287328441C97931131 @default.
- W4287328441 hasLocation W42873284411 @default.
- W4287328441 hasOpenAccess W4287328441 @default.
- W4287328441 hasPrimaryLocation W42873284411 @default.
- W4287328441 hasRelatedWork W2047094413 @default.
- W4287328441 hasRelatedWork W2128201184 @default.
- W4287328441 hasRelatedWork W2134501463 @default.
- W4287328441 hasRelatedWork W2137675115 @default.
- W4287328441 hasRelatedWork W2141540671 @default.
- W4287328441 hasRelatedWork W2143620265 @default.
- W4287328441 hasRelatedWork W2843351054 @default.
- W4287328441 hasRelatedWork W3131215311 @default.
- W4287328441 hasRelatedWork W3160939593 @default.
- W4287328441 hasRelatedWork W4287328441 @default.
- W4287328441 isParatext "false" @default.
- W4287328441 isRetracted "false" @default.
- W4287328441 workType "article" @default.