1 |
Hippocampal and auditory contributions to speech segmentation
|
|
|
|
In: ISSN: 0010-9452 ; Cortex ; https://hal.archives-ouvertes.fr/hal-03604957 ; Cortex, Elsevier, 2022, ⟨10.1016/j.cortex.2022.01.017⟩ (2022)
|
|
BASE
|
|
Show details
|
|
2 |
Speaking clearly improves speech segmentation by statistical learning under optimal listening conditions ...
|
|
|
|
BASE
|
|
Show details
|
|
3 |
The effect of lengthening aspiration on speech segmentation ...
|
|
|
|
BASE
|
|
Show details
|
|
4 |
End-to-end speaker segmentation for overlap-aware resegmentation
|
|
|
|
In: Interspeech 2021 ; https://hal-univ-lemans.archives-ouvertes.fr/hal-03257524 ; Interspeech 2021, Aug 2021, Brno, Czech Republic ; https://www.interspeech2021.org/ (2021)
|
|
BASE
|
|
Show details
|
|
5 |
Impact of Encoding and Segmentation Strategies on End-to-End Simultaneous Speech Translation
|
|
|
|
In: INTERSPEECH 2021 ; https://hal.archives-ouvertes.fr/hal-03372487 ; INTERSPEECH 2021, Aug 2021, Brno, Czech Republic (2021)
|
|
BASE
|
|
Show details
|
|
6 |
Oscillatory activity and EEG phase synchrony of concurrent word segmentation and meaning-mapping in 9-year-old children
|
|
|
|
In: ISSN: 1878-9293 ; EISSN: 1878-9307 ; Developmental Cognitive Neuroscience ; https://hal.archives-ouvertes.fr/hal-03334735 ; Developmental Cognitive Neuroscience, Elsevier, 2021, 51, pp.101010. ⟨10.1016/j.dcn.2021.101010⟩ (2021)
|
|
BASE
|
|
Show details
|
|
7 |
Transdisciplinary Analysis of a Corpus of French Newsreels: The ANTRACT Project
|
|
|
|
In: ISSN: 1938-4122 ; Digital Humanities Quarterly ; https://hal.archives-ouvertes.fr/hal-03166755 ; Digital Humanities Quarterly, Alliance of Digital Humanities, 2021, Special Issue on AudioVisual Data in DH, 15 (1) ; http://digitalhumanities.org/dhq/ (2021)
|
|
BASE
|
|
Show details
|
|
8 |
Speaking clearly improves speech segmentation by statistical learning under optimal listening conditions
|
|
|
|
In: Laboratory Phonology: Journal of the Association for Laboratory Phonology; Vol 12, No 1 (2021); 14 ; 1868-6354 (2021)
|
|
BASE
|
|
Show details
|
|
11 |
Production of nonce words to establish the cues for prominence and grouping in English ...
|
|
|
|
BASE
|
|
Show details
|
|
12 |
Early Tashelhiyt Berber word segmentation: the role of the Possible Word Constraint ...
|
|
|
|
BASE
|
|
Show details
|
|
14 |
The Iambic Trochaic Law in speech: The case of Japanese ...
|
|
|
|
BASE
|
|
Show details
|
|
16 |
Developing Core Technologies for Resource-Scarce Nguni Languages
|
|
|
|
In: Information; Volume 12; Issue 12; Pages: 520 (2021)
|
|
BASE
|
|
Show details
|
|
17 |
Discovering structure in speech recordings: Unsupervised learning of word and phoneme like units for automatic speech recognition
|
|
|
|
In: Fraunhofer IAIS (2021)
|
|
Abstract:
While speech recordings are easy to obtain, the transcription of those recordings can be very costly and time-consuming. Therefore, automatic methods to derive such transcriptions from unlabeled data can help simplifying the training of speech recognizers in languages where little to no labeled training data is available. This thesis investigates and introduces methods to automatically learn transcriptions from audio recordings only. Algorithms for the unsupervised learning of phonemes, the smallest units in speech, and words are presented. These methods can then be used for the automatic training of a speech recognizer from unlabeled data. This thesis investigates these unsupervised learning methods separately for the learning of phonemes and words. The main focus of this thesis is laid on the unsupervised learning of words in hierarchical models consisting of phoneme and word transcriptions. Three main approaches are investigated. Firstly, heuristic methods. Secondly, two variants of statistical model-based approaches. The first variant is based on a probabilistic pronunciation lexicon while the second approach is based on word segmentation over lattices, instead of a single best sequence. Finally, a fully unsupervised system with unsupervised phoneme discovery and unsupervised word segmentation combined, is presented. The thesis concludes by integrating the unsupervised phoneme and word discovery into a semantic inference task in the setting of a simple command and control interface to demonstrate the usability of unsupervised learned phonemes and words in upstream tasks and their ability to improve their performance over purely supervised methods.
|
|
Keyword:
Acoustic Unit Discovery; ASR; automatic speech recognition; unsupervised learning; Unsupervised Word Segmentation
|
|
URL: http://publica.fraunhofer.de/documents/N-644770.html https://doi.org/10.17619/UNIPB/1-1252
|
|
BASE
|
|
Hide details
|
|
18 |
Projecting action spaces. On the interactional relevance of cesural areas in co-enactments
|
|
|
|
In: Open Linguistics, Vol 7, Iss 1, Pp 638-665 (2021) (2021)
|
|
BASE
|
|
Show details
|
|
19 |
Streaming cascade-based speech translation leveraged by a direct segmentation model
|
|
|
|
BASE
|
|
Show details
|
|
20 |
Developing Resources for Automated Speech Processing of Quebec French
|
|
|
|
In: Proceedings of the 12th Language Resources and Evaluation Conference ; 12th Language Resources and Evaluation Conference ; https://hal.archives-ouvertes.fr/hal-03042864 ; 12th Language Resources and Evaluation Conference, 2020, marseille, France. pp.5323-5328 (2020)
|
|
BASE
|
|
Show details
|
|
|
|