6 articles tagged with #speech-to-text. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.
Researchers introduce ESRT, a privacy-preserving edge-cloud framework for multilingual speech-to-text translation that processes voice data locally while transmitting only compressed features to the cloud. The system achieves state-of-the-art performance across 45 languages while reducing bandwidth requirements by 10x and preventing voiceprint leakage.
Researchers find that cross-attention mechanisms in speech-to-text models only explain about 50% of how the decoder attends to input, contradicting widespread assumptions that attention scores reliably indicate which parts of the audio are most relevant. The study across multiple model scales shows attention provides an incomplete view of the factors driving predictions.
Researchers introduce ELF-S2T, a novel continuous-target generative model for speech-to-text tasks that combines audio conditioning with diffusion-based language modeling. The approach achieves competitive performance on ASR and speech translation while revealing that both tasks share common error patterns rooted in continuous latent space representations.
Researchers introduce DLLM-VSR, a diffusion-based large language model framework for visual speech recognition that replaces traditional left-to-right decoding with iterative masked denoising. The system achieves state-of-the-art 19.5% word error rate on LRS3 by using confidence-based unmasking and length-guided candidate decoding to resolve visual ambiguities.
Google has released MedGemma 1.5 for next-generation medical image interpretation and MedASR for medical speech-to-text applications. These new AI tools represent significant advancements in healthcare AI capabilities, focusing on specialized medical applications.
Google has quietly launched a new offline-first AI dictation app for iOS that utilizes Gemma AI models. The app appears to be positioning itself as a competitor to existing dictation solutions like Wispr Flow by offering offline functionality.
