Sökning: "wav2vec2"

Hittade 5 uppsatser innehållade ordet wav2vec2.

  1. 1. Speech Classification using Acoustic embedding and Large Language Models Applied on Alzheimer’s Disease Prediction Task

    Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

    Författare :Maryam Kheirkhahzadeh; [2023]
    Nyckelord :Speech classification; Alzheimer’s disease detection; GPT-3; BERT; Text embedding; Dementia; wav2vec2.0; Klassificering av tal; detektion av Alzheimer’s sjukdom; GPT-3; BERT; textinbäddning; demens; wav2vec2.0;

    Sammanfattning : Alzheimer’s sjukdom är en neurodegenerativ sjukdom som leder till demens. Den kan börja tyst i de tidiga stadierna och fortsätta under åren till en allvarlig och obotlig fas. Språkstörningar uppstår ofta som ett av de tidiga symptomen och kan till slut leda till fullständig mutism i de avancerade stadierna av sjukdomen. LÄS MER

  3. 2. Improving accuracy of speech recognition for low resource accents : Testing the performance of fine-tuned Wav2vec2 models on accented Swedish

    Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

    Författare :Arash Dabiri; [2023]
    Nyckelord :Speech-to-text; deep learning; accents; wav2vec; tal-till-text; djupinlärning; brytningar; wav2vec;

    Sammanfattning : While the field of speech recognition has recently advanced quickly, even the highest performing models struggle with accents. There are several methods of improving the performance on accents, but many are hard to implement or need high amounts of data and are therefore costly to implement. LÄS MER

  4. 3. Domain Adaptation with N-gram Language Models for Swedish Automatic Speech Recognition : Using text data augmentation to create domain-specific n-gram models for a Swedish open-source wav2vec 2.0 model

    Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

    Författare :Viktor Enzell; [2022]
    Nyckelord :Automatic Speech Recognition; Domain Adaptation; Language Models; Ngram Models; Wav2vec2; Taligenkänning; Domänanpassning; Språkmodeller; N-gramModeller; Wav2vec2;

    Sammanfattning : Automatic Speech Recognition (ASR) enables a wide variety of practical applications. However, many applications have their own domain-specific words, creating a gap between training and test data when used in practice. LÄS MER

  5. 4. Automatic Annotation of Speech: Exploring Boundaries within Forced Alignment for Swedish and Norwegian

    Master-uppsats, Uppsala universitet/Institutionen för lingvistik och filologi

    Författare :Klaudia Biczysko; [2022]
    Nyckelord :forced alignment; automatic speech recognition; ASR; natural language processing; under-resourced languages; Swedish; Norwegian; CTC segmentation; wav2vec2; kaldi; HTK; dynamic time warping;

    Sammanfattning : In Automatic Speech Recognition, there is an extensive need for time-aligned data. Manual speech segmentation has been shown to be more laborious than manual transcription, especially when dealing with tens of hours of speech. LÄS MER

  7. 5. Automatic Speech Recognition for low-resource languages using Wav2Vec2 : Modern Standard Arabic (MSA) as an example of a low-resource language

    Master-uppsats, Högskolan Dalarna/Institutionen för information och teknik

    Författare :Taha Zouhair; [2021]
    Nyckelord :Automatic Speech Recognition; Facebook Wav2Vec; Mozilla Common Voice; Low-Resource Language;

    Sammanfattning : The need for fully automatic translation at DigitalTolk, a Stockholm-based company providing translation services, leads to exploring Automatic Speech Recognition as a first step for Modern Standard Arabic (MSA). Facebook AI recently released a second version of its Wav2Vec models, dubbed Wav2Vec 2. LÄS MER