Statistical language and speech processing : 9th international conference, SLSP 2021, Cardiff, UK, November 23-25, 2021 : proceedings / / edited by Luis Espinosa-Anke, Carlos Martín-Vide, Irena Spasić

• Pubblicazione: Cham, Switzerland : , : Springer, , [2021]; ©2021
• Descrizione fisica: 1 online resource (119 pages)
• Disciplina: 006.35
• Soggetto topico: Speech processing systems
• Persona (resp. second.): Espinosa-AnkeLuis; Martín VideCarlos; SpasićIrena
• Nota di bibliografia: Includes bibliographical references and index.
• Nota di contenuto: Intro -- Preface -- Organization -- Contents -- Language -- Improving German Image Captions Using Machine Translation and Transfer Learning -- 1 Related Work -- 2 Method -- 2.1 Image Captioning Datasets -- 2.2 Image Captioning Model -- 2.3 Caption Generation Methods -- 3 Evaluation -- 3.1 Metrics -- 3.2 Hypothesis -- 3.3 Results -- 4 Discussion -- 5 Conclusion -- References -- Automatic News Article Generation from Legislative Proceedings: A Phenom-Based Approach -- 1 Introduction and Motivation -- 1.1 Motivation -- 1.2 Organization -- 2 Related Work -- 3 Approach and Development -- 3.1 Phenom System -- 3.2 Illustrative Phenom Examples -- 3.3 Template-Based Text Generation and Planning -- 4 Research Study -- 4.1 Design and Protocol -- 4.2 Results and Discussion -- 5 Conclusion and Future Work -- References -- Comparison of Czech Transformers on Text Classification Tasks -- 1 Introduction -- 2 Related Work -- 3 Datasets -- 3.1 Pre-training Datasets -- 3.2 Text Classification Datasets -- 4 Models -- 4.1 Pre-training -- 4.2 Fine-Tuning -- 4.3 Evaluation -- 5 Results -- 6 Conclusions -- References -- Constructing Sentiment Lexicon with Game for Annotation Collection -- 1 Introduction -- 2 Related Work -- 2.1 Games with a Purpose -- 2.2 Sentiment Analysis in Slovak -- 3 Game for Sentiment Collection -- 4 Experiments and Evaluation -- 4.1 Sentiment Lexicon -- 5 Conclusion -- References -- Robustness of Named Entity Recognition: Case of Latvian -- 1 Introduction -- 2 Data -- 3 Models -- 3.1 Impact of Error Types on NER -- 3.2 Data Augmentation and Robustness -- 3.3 Finding a Robust Model -- 4 Conclusions -- References -- Speech -- Use of Speaker Metadata for Improving Automatic Pronunciation Assessment -- 1 Introduction -- 2 The ASR Approach for Automatic Pronunciation Assessment -- 3 A Segment Based Approach for Mispronunciation Detection. ; 4 Attention Based Model -- 5 The Data -- 6 Encoding of Speaker Factors -- 7 Experiments -- 8 Results and Discussion -- 9 Conclusions -- References -- Augmenting ASR for User-Generated Videos with Semi-supervised Training and Acoustic Model Adaptation for Spoken Content Retrieval -- 1 Introduction -- 2 Semi-supervised Acoustic and Language Modelling -- 3 Adaptation of Acoustic Model Using Content Genre -- 4 ASR Experiments -- 4.1 Creation of Manual Transcripts for Blip10000 -- 4.2 Experimental Setup -- 4.3 Experimental Results -- 5 SCR Experiments -- 5.1 Creation of Known-Item Queries for Blip10000 -- 5.2 Experimental Setup -- 5.3 Experimental Results and Analysis -- 6 Conclusions and Further Work -- References -- Various DNN-HMM Architectures Used in Acoustic Modeling with Single-Speaker and Single-Channel -- 1 Introduction -- 2 Training and Test Data Set -- 3 Experimental Setup -- 3.1 Acoustic Feature Extraction -- 3.2 Acoustic Modeling -- 3.3 Language Modeling -- 3.4 Decoding -- 4 Experiments -- 4.1 HMM-topology and Context-Dependent Modeling -- 4.2 DNN Context -- 4.3 Implementation Issues -- 4.4 Amount of Training Data -- 5 Conclusion -- References -- Invariant Representation Learning for Robust Far-Field Speaker Recognition -- 1 Introduction -- 2 Base Model Architecture -- 2.1 Encoder Network: ResNet34 -- 2.2 Classification Layer: Additive Margin Softmax (AM-Softmax) -- 2.3 Back-End -- 3 Invariant Representation Learning (IRL) -- 3.1 Text-Dependent IRL (TD-IRL) -- 3.2 Text Independent IRL (TI-IRL) -- 3.3 Deep Features IRL (DF-IRL) -- 4 Experimental Setup -- 4.1 VOiCES Dataset -- 4.2 Base Model Training -- 4.3 Data Split for Evaluation -- 5 Results -- 6 Conclusions -- References -- Author Index.
• Titolo autorizzato: Statistical language and speech processing
• ISBN: 3-030-89579-3
• Formato: Materiale a stampa
• Livello bibliografico: Monografia
• Lingua di pubblicazione: Inglese
• Record Nr.: 996464381903316
• Lo trovi qui: Univ. di Salerno
• Serie: Lecture Notes in Computer Science