0% Complete
English
صفحه اصلی
/
پانزدهمین کنفرانس بین المللی فناوری اطلاعات و دانش
Embedded speech encoder for low-resource languages
نویسندگان :
Alireza A.Tabatabaei
1
Pouria Sameti
2
Ali Bohlooli
3
1- University of Isfahan
2- University of Isfahan
3- University of Isfahan
کلمات کلیدی :
Embedded Systems،Embedded AI،Embedded Speech embedding
چکیده :
Although high-performance artificial intelligence (AI) models require substantial computational resources, embedded systems are constrained by limited hardware capabilities, such as memory and processing power. On the other hand, embedded systems have a broad range of applications, making the integration of AI and embedded systems a prominent topic in both hardware and AI research. Creating powerful speech embeddings for embedded systems is challenging, as such models, like Wave2Vec, are typically computationally intensive. Additionally, the scarcity of data for many low-resource languages further complicates the development of high-performance models. To address these challenges, we utilized BERT to generate speech embeddings. BERT was selected because, in addition to producing meaningful embeddings, it is trained on numerous low-resource languages and facilitates the design of efficient decoders. This study introduces a compact speech encoder tailored for low-resource languages, capable of functioning as an encoder across a diverse range of speech tasks. To achieve this, we utilized BERT to generate meaningful embeddings. However, due to the high dimensionality of BERT embeddings, which imposes significant computational demands on many embedded systems, we applied dimensionality reduction techniques. The reduced-dimensional vectors were subsequently used as labels for speech data to train a model composed of convolutional neural networks (CNNs) and fully connected layers. Finally, we demonstrated the encoder's effectiveness through an application in speech command recognition.
لیست مقالات
لیست مقالات بایگانی شده
BMPA- DSL: Binary Marine Predators Algorithm to Identify Driver's Different Levels of Stress
Mahtab Vaezi - Mehdi Nasri - Farhad Azimifar - Mahdi Mosleh
Blockchain-based Secure UAV-assisted Battlefield Operation underlying 5G
Dhruvi Pancholi - Nilesh Kumar Jadav - Sudeep Tanwar - Deepak Garg - S. Mohammadali Zanjani
User Preferences Elicitation in Bilateral Automated Negotiation Using Recursive Least Square Estimation
Farnaz Salmanian - Dr Hamid Jazayeri - Dr Javad Kazemitabar
Sentiment Analysis of the Amazon Customers Using the BiGRU Neural Network Enhanced by Attention Mechanism
Sara Sinan Salman al-Abedi - Keyvan Mohebbi
Classical-Quantum Multiple Access Wiretap Channel with Common Message: One-shot Rate Region
Hadi Aghaee - Dr Bahareh Akhbari
بهبود دقت و کارایی در شبکههای عصبی کانولوشنی با استفاده از روشهای محاسبات تقریبی
محمدرضا رفیعی نژاد - محمدرضا بینش مروستی - سید امیر اصغری
Architectural Insights: Comparing Weight Stationary and Output Stationary Systolic Arrays for Efficient Computation
Mahdi Kalbasi
Binary water stream algorithm: a new meta-heuristic optimization technique
Faezeh Rahimi Sebdani - Mehdi Nasri
یک روش خوشه بندی گره ها برای شبکه های حسگر بیسیم با هدف بهبود متوازن سازی بار مبتنی بر تکنیک تاپسیس
راضیه حسین رضایی - فهیمه یزدان پناه
حفظ حریم خصوصی در انتشار نسخه های متوالی دادههای شبکه اجتماعی با امکان افزایش یال
طاهره سرزهی - دکتر مهری رجایی طاهره سرزهی - مهری رجایی -
بیشتر
ثمین همایش، سامانه مدیریت کنفرانس ها و جشنواره ها - نگارش 42.0.3