Matches in SemOpenAlex for { <https://semopenalex.org/work/W4319862404> ?p ?o ?g. }
Showing items 1 to 100 of
100
with 100 items per page.
- W4319862404 abstract "Compressing self-supervised models has become increasingly necessary, as self-supervised models become larger. While previous approaches have primarily focused on compressing the model size, shortening sequences is also effective in reducing the computational cost. In this work, we study fixed-length and variable-length subsampling along the time axis in self-supervised learning. We explore how individual downstream tasks are sensitive to input frame rates. Subsampling while training self-supervised models not only improves the overall performance on downstream tasks under certain frame rates, but also brings significant speed-up in inference. Variable-length subsampling performs particularly well under low frame rates. In addition, if we have access to phonetic boundaries, we find no degradation in performance for an average frame rate as low as 10 Hz." @default.
- W4319862404 created "2023-02-11" @default.
- W4319862404 creator A5001291873 @default.
- W4319862404 creator A5008336983 @default.
- W4319862404 creator A5030898980 @default.
- W4319862404 creator A5040508737 @default.
- W4319862404 creator A5059858850 @default.
- W4319862404 creator A5062573067 @default.
- W4319862404 creator A5067668991 @default.
- W4319862404 date "2023-01-09" @default.
- W4319862404 modified "2023-09-24" @default.
- W4319862404 title "On Compressing Sequences for Self-Supervised Speech Models" @default.
- W4319862404 cites W1494198834 @default.
- W4319862404 cites W2010291496 @default.
- W4319862404 cites W2190506272 @default.
- W4319862404 cites W2296681920 @default.
- W4319862404 cites W2327501763 @default.
- W4319862404 cites W2526425061 @default.
- W4319862404 cites W2530876040 @default.
- W4319862404 cites W2627092829 @default.
- W4319862404 cites W2747874407 @default.
- W4319862404 cites W2802023636 @default.
- W4319862404 cites W2916113431 @default.
- W4319862404 cites W2933138175 @default.
- W4319862404 cites W2962784628 @default.
- W4319862404 cites W2962826786 @default.
- W4319862404 cites W2963211739 @default.
- W4319862404 cites W2963250244 @default.
- W4319862404 cites W3016167541 @default.
- W4319862404 cites W3197974236 @default.
- W4319862404 cites W3198134274 @default.
- W4319862404 cites W3198782837 @default.
- W4319862404 cites W3200129129 @default.
- W4319862404 cites W3203140070 @default.
- W4319862404 cites W3209059054 @default.
- W4319862404 cites W3209984917 @default.
- W4319862404 cites W4297841557 @default.
- W4319862404 cites W4297841794 @default.
- W4319862404 doi "https://doi.org/10.1109/slt54892.2023.10022991" @default.
- W4319862404 hasPublicationYear "2023" @default.
- W4319862404 type Work @default.
- W4319862404 citedByCount "0" @default.
- W4319862404 crossrefType "proceedings-article" @default.
- W4319862404 hasAuthorship W4319862404A5001291873 @default.
- W4319862404 hasAuthorship W4319862404A5008336983 @default.
- W4319862404 hasAuthorship W4319862404A5030898980 @default.
- W4319862404 hasAuthorship W4319862404A5040508737 @default.
- W4319862404 hasAuthorship W4319862404A5059858850 @default.
- W4319862404 hasAuthorship W4319862404A5062573067 @default.
- W4319862404 hasAuthorship W4319862404A5067668991 @default.
- W4319862404 hasBestOaLocation W43198624042 @default.
- W4319862404 hasConcept C119857082 @default.
- W4319862404 hasConcept C126042441 @default.
- W4319862404 hasConcept C134306372 @default.
- W4319862404 hasConcept C136389625 @default.
- W4319862404 hasConcept C153180895 @default.
- W4319862404 hasConcept C154945302 @default.
- W4319862404 hasConcept C182365436 @default.
- W4319862404 hasConcept C2776214188 @default.
- W4319862404 hasConcept C28490314 @default.
- W4319862404 hasConcept C3261483 @default.
- W4319862404 hasConcept C33923547 @default.
- W4319862404 hasConcept C41008148 @default.
- W4319862404 hasConcept C50644808 @default.
- W4319862404 hasConcept C76155785 @default.
- W4319862404 hasConceptScore W4319862404C119857082 @default.
- W4319862404 hasConceptScore W4319862404C126042441 @default.
- W4319862404 hasConceptScore W4319862404C134306372 @default.
- W4319862404 hasConceptScore W4319862404C136389625 @default.
- W4319862404 hasConceptScore W4319862404C153180895 @default.
- W4319862404 hasConceptScore W4319862404C154945302 @default.
- W4319862404 hasConceptScore W4319862404C182365436 @default.
- W4319862404 hasConceptScore W4319862404C2776214188 @default.
- W4319862404 hasConceptScore W4319862404C28490314 @default.
- W4319862404 hasConceptScore W4319862404C3261483 @default.
- W4319862404 hasConceptScore W4319862404C33923547 @default.
- W4319862404 hasConceptScore W4319862404C41008148 @default.
- W4319862404 hasConceptScore W4319862404C50644808 @default.
- W4319862404 hasConceptScore W4319862404C76155785 @default.
- W4319862404 hasFunder F4320307764 @default.
- W4319862404 hasFunder F4320309327 @default.
- W4319862404 hasFunder F4320310145 @default.
- W4319862404 hasFunder F4320310598 @default.
- W4319862404 hasLocation W43198624041 @default.
- W4319862404 hasLocation W43198624042 @default.
- W4319862404 hasOpenAccess W4319862404 @default.
- W4319862404 hasPrimaryLocation W43198624041 @default.
- W4319862404 hasRelatedWork W2511279186 @default.
- W4319862404 hasRelatedWork W2979403462 @default.
- W4319862404 hasRelatedWork W2981850339 @default.
- W4319862404 hasRelatedWork W3094076422 @default.
- W4319862404 hasRelatedWork W3095538999 @default.
- W4319862404 hasRelatedWork W3162567751 @default.
- W4319862404 hasRelatedWork W3210156800 @default.
- W4319862404 hasRelatedWork W4221088574 @default.
- W4319862404 hasRelatedWork W4226172683 @default.
- W4319862404 hasRelatedWork W4249546094 @default.
- W4319862404 isParatext "false" @default.
- W4319862404 isRetracted "false" @default.
- W4319862404 workType "article" @default.