Matches in SemOpenAlex for { <https://semopenalex.org/work/W3089220633> ?p ?o ?g. }
- W3089220633 abstract "Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a key component of spoken dialog systems. In this work, we propose a new approach to learn generic representations adapted to spoken dialog, which we evaluate on a new benchmark we call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (texttt{SILICONE}). texttt{SILICONE} is model-agnostic and contains 10 different datasets of various sizes. We obtain our representations with a hierarchical encoder based on transformer architectures, for which we extend two well-known pre-training objectives. Pre-training is performed on OpenSubtitles: a large corpus of spoken dialog containing over $2.3$ billion of tokens. We demonstrate how hierarchical encoders achieve competitive results with consistently fewer parameters compared to state-of-the-art models and we show their importance for both pre-training and fine-tuning." @default.
- W3089220633 created "2020-10-01" @default.
- W3089220633 creator A5005561269 @default.
- W3089220633 creator A5027506291 @default.
- W3089220633 creator A5044425827 @default.
- W3089220633 creator A5055866824 @default.
- W3089220633 creator A5064876801 @default.
- W3089220633 date "2020-09-23" @default.
- W3089220633 modified "2023-09-30" @default.
- W3089220633 title "Hierarchical Pre-training for Sequence Labelling in Spoken Dialog" @default.
- W3089220633 cites W1488153692 @default.
- W3089220633 cites W1503312748 @default.
- W3089220633 cites W1522301498 @default.
- W3089220633 cites W2095705004 @default.
- W3089220633 cites W2124246394 @default.
- W3089220633 cites W2125320996 @default.
- W3089220633 cites W2128970689 @default.
- W3089220633 cites W2132555391 @default.
- W3089220633 cites W2143023047 @default.
- W3089220633 cites W2153579005 @default.
- W3089220633 cites W2155939513 @default.
- W3089220633 cites W2165644552 @default.
- W3089220633 cites W2166637769 @default.
- W3089220633 cites W2250539671 @default.
- W3089220633 cites W2296712013 @default.
- W3089220633 cites W2401527985 @default.
- W3089220633 cites W2419539795 @default.
- W3089220633 cites W2462831000 @default.
- W3089220633 cites W2525778437 @default.
- W3089220633 cites W2624871570 @default.
- W3089220633 cites W2741802726 @default.
- W3089220633 cites W2761590056 @default.
- W3089220633 cites W2787560479 @default.
- W3089220633 cites W2788420618 @default.
- W3089220633 cites W2792643794 @default.
- W3089220633 cites W2799054028 @default.
- W3089220633 cites W2806412155 @default.
- W3089220633 cites W2890538051 @default.
- W3089220633 cites W2891732163 @default.
- W3089220633 cites W2892068161 @default.
- W3089220633 cites W2895615590 @default.
- W3089220633 cites W2896439639 @default.
- W3089220633 cites W2899663614 @default.
- W3089220633 cites W2900338016 @default.
- W3089220633 cites W2911109671 @default.
- W3089220633 cites W2939365567 @default.
- W3089220633 cites W2943766373 @default.
- W3089220633 cites W2945918281 @default.
- W3089220633 cites W2947972709 @default.
- W3089220633 cites W2948902769 @default.
- W3089220633 cites W2949433733 @default.
- W3089220633 cites W2950541952 @default.
- W3089220633 cites W2951612168 @default.
- W3089220633 cites W2962854379 @default.
- W3089220633 cites W2963168538 @default.
- W3089220633 cites W2963341956 @default.
- W3089220633 cites W2963403868 @default.
- W3089220633 cites W2963550483 @default.
- W3089220633 cites W2963765493 @default.
- W3089220633 cites W2965373594 @default.
- W3089220633 cites W2970597249 @default.
- W3089220633 cites W2971222961 @default.
- W3089220633 cites W2973230496 @default.
- W3089220633 cites W2979364224 @default.
- W3089220633 cites W2979884483 @default.
- W3089220633 cites W2990568760 @default.
- W3089220633 cites W2994980856 @default.
- W3089220633 cites W2996428491 @default.
- W3089220633 cites W3005957694 @default.
- W3089220633 cites W3008671257 @default.
- W3089220633 cites W3036288367 @default.
- W3089220633 cites W3087841754 @default.
- W3089220633 cites W3105966348 @default.
- W3089220633 cites W362060233 @default.
- W3089220633 cites W647206835 @default.
- W3089220633 cites W2970603031 @default.
- W3089220633 cites W3013040126 @default.
- W3089220633 hasPublicationYear "2020" @default.
- W3089220633 type Work @default.
- W3089220633 sameAs 3089220633 @default.
- W3089220633 citedByCount "3" @default.
- W3089220633 countsByYear W30892206332020 @default.
- W3089220633 countsByYear W30892206332021 @default.
- W3089220633 crossrefType "posted-content" @default.
- W3089220633 hasAuthorship W3089220633A5005561269 @default.
- W3089220633 hasAuthorship W3089220633A5027506291 @default.
- W3089220633 hasAuthorship W3089220633A5044425827 @default.
- W3089220633 hasAuthorship W3089220633A5055866824 @default.
- W3089220633 hasAuthorship W3089220633A5064876801 @default.
- W3089220633 hasConcept C111919701 @default.
- W3089220633 hasConcept C118505674 @default.
- W3089220633 hasConcept C121332964 @default.
- W3089220633 hasConcept C13280743 @default.
- W3089220633 hasConcept C136764020 @default.
- W3089220633 hasConcept C144024400 @default.
- W3089220633 hasConcept C154945302 @default.
- W3089220633 hasConcept C162324750 @default.
- W3089220633 hasConcept C165801399 @default.
- W3089220633 hasConcept C173853756 @default.
- W3089220633 hasConcept C185798385 @default.