A multimodal dataset of over 100,000 structured examples designed for training models that translate biosignals into natural-language communication. The dataset includes eye-tracking patterns, EMG signals, facial expression indicators, multimodal combinations, contextual metadata, urgency levels, and target natural-language outputs.
This dataset was created to support research in assistive communication, especially for… See the full description on the dataset page:
https://huggingface.co/datasets/0xroyce/silent-voice-100k.
