PhonoBiEmbedNet: A Phoneme and Bigram Embedding Framework for Low-Resource Spoken Word Recognition

Key Points

• Enhances spoken word recognition using phoneme and bigram embeddings, leading to improved accuracy.
• Achieves significant performance improvements, reporting a 20% increase in recognition rates on low-resource datasets.
• Harnesses deep learning techniques in developing the phoneme and bigram embedding framework for efficient processing.
• Societal implications suggest better voice recognition technologies could support diverse languages and dialects.