Matches in SemOpenAlex for { <https://semopenalex.org/work/W3213804535> ?p ?o ?g. }
- W3213804535 endingPage "176" @default.
- W3213804535 startingPage "163" @default.
- W3213804535 abstract "The distributed and continuous representations used by neural networks are at odds with representations employed in linguistics, which are typically symbolic. Vector quantization has been proposed as a way to induce discrete neural representations that are closer in nature to their linguistic counterparts. However, it is not clear which metrics are the best-suited to analyze such discrete representations. We compare the merits of four commonly used metrics in the context of weakly supervised models of spoken language. We compare the results they show when applied to two different models, while systematically studying the effect of the placement and size of the discretization layer. We find that different evaluation regimes can give inconsistent results. While we can attribute them to the properties of the different metrics in most cases, one point of concern remains: the use of minimal pairs of phoneme triples as stimuli disadvantages larger discrete unit inventories, unlike metrics applied to complete utterances. Furthermore, while in general vector quantization induces representations that correlate with units posited in linguistics, the strength of this correlation is only moderate." @default.
- W3213804535 created "2021-11-22" @default.
- W3213804535 creator A5000074624 @default.
- W3213804535 creator A5020470786 @default.
- W3213804535 creator A5022698890 @default.
- W3213804535 creator A5047382001 @default.
- W3213804535 date "2021-11-01" @default.
- W3213804535 modified "2023-09-27" @default.
- W3213804535 title "Discrete representations in neural models of spoken language" @default.
- W3213804535 cites W1524333225 @default.
- W3213804535 cites W2024081693 @default.
- W3213804535 cites W2119775030 @default.
- W3213804535 cites W2138615112 @default.
- W3213804535 cites W2194775991 @default.
- W3213804535 cites W2242818861 @default.
- W3213804535 cites W2786608204 @default.
- W3213804535 cites W2888912391 @default.
- W3213804535 cites W2946296745 @default.
- W3213804535 cites W2949079242 @default.
- W3213804535 cites W2962776659 @default.
- W3213804535 cites W2962862718 @default.
- W3213804535 cites W2963430224 @default.
- W3213804535 cites W2963620343 @default.
- W3213804535 cites W2963799213 @default.
- W3213804535 cites W2963902314 @default.
- W3213804535 cites W2964054038 @default.
- W3213804535 cites W2964121744 @default.
- W3213804535 cites W2964204621 @default.
- W3213804535 cites W2971709506 @default.
- W3213804535 cites W2972374322 @default.
- W3213804535 cites W2973047874 @default.
- W3213804535 cites W2984673553 @default.
- W3213804535 cites W2995680346 @default.
- W3213804535 cites W3035305735 @default.
- W3213804535 cites W3035750922 @default.
- W3213804535 cites W3087357110 @default.
- W3213804535 cites W3095361818 @default.
- W3213804535 cites W3096196861 @default.
- W3213804535 cites W3096216486 @default.
- W3213804535 cites W3097286738 @default.
- W3213804535 cites W3098952151 @default.
- W3213804535 cites W3100813302 @default.
- W3213804535 cites W3105148948 @default.
- W3213804535 cites W3157861865 @default.
- W3213804535 cites W3171345413 @default.
- W3213804535 cites W2593779438 @default.
- W3213804535 hasPublicationYear "2021" @default.
- W3213804535 type Work @default.
- W3213804535 sameAs 3213804535 @default.
- W3213804535 citedByCount "0" @default.
- W3213804535 crossrefType "journal-article" @default.
- W3213804535 hasAuthorship W3213804535A5000074624 @default.
- W3213804535 hasAuthorship W3213804535A5020470786 @default.
- W3213804535 hasAuthorship W3213804535A5022698890 @default.
- W3213804535 hasAuthorship W3213804535A5047382001 @default.
- W3213804535 hasConcept C11413529 @default.
- W3213804535 hasConcept C134306372 @default.
- W3213804535 hasConcept C137293760 @default.
- W3213804535 hasConcept C154945302 @default.
- W3213804535 hasConcept C199833920 @default.
- W3213804535 hasConcept C204321447 @default.
- W3213804535 hasConcept C2524010 @default.
- W3213804535 hasConcept C28490314 @default.
- W3213804535 hasConcept C28719098 @default.
- W3213804535 hasConcept C28855332 @default.
- W3213804535 hasConcept C33923547 @default.
- W3213804535 hasConcept C41008148 @default.
- W3213804535 hasConcept C50644808 @default.
- W3213804535 hasConcept C73000952 @default.
- W3213804535 hasConceptScore W3213804535C11413529 @default.
- W3213804535 hasConceptScore W3213804535C134306372 @default.
- W3213804535 hasConceptScore W3213804535C137293760 @default.
- W3213804535 hasConceptScore W3213804535C154945302 @default.
- W3213804535 hasConceptScore W3213804535C199833920 @default.
- W3213804535 hasConceptScore W3213804535C204321447 @default.
- W3213804535 hasConceptScore W3213804535C2524010 @default.
- W3213804535 hasConceptScore W3213804535C28490314 @default.
- W3213804535 hasConceptScore W3213804535C28719098 @default.
- W3213804535 hasConceptScore W3213804535C28855332 @default.
- W3213804535 hasConceptScore W3213804535C33923547 @default.
- W3213804535 hasConceptScore W3213804535C41008148 @default.
- W3213804535 hasConceptScore W3213804535C50644808 @default.
- W3213804535 hasConceptScore W3213804535C73000952 @default.
- W3213804535 hasLocation W32138045351 @default.
- W3213804535 hasOpenAccess W3213804535 @default.
- W3213804535 hasPrimaryLocation W32138045351 @default.
- W3213804535 hasRelatedWork W1818219649 @default.
- W3213804535 hasRelatedWork W1937075317 @default.
- W3213804535 hasRelatedWork W2141599568 @default.
- W3213804535 hasRelatedWork W2251012068 @default.
- W3213804535 hasRelatedWork W2286410738 @default.
- W3213804535 hasRelatedWork W2404733436 @default.
- W3213804535 hasRelatedWork W2418478089 @default.
- W3213804535 hasRelatedWork W2557321750 @default.
- W3213804535 hasRelatedWork W2786382385 @default.
- W3213804535 hasRelatedWork W2897058755 @default.
- W3213804535 hasRelatedWork W2946817437 @default.
- W3213804535 hasRelatedWork W2949536053 @default.