Researchers release PulseLM, a large-scale PPG-text dataset for physiological inference

Researchers have introduced PulseLM, a novel dataset and benchmark designed for learning the relationship between photoplethysmography (PPG) signals and text. This dataset integrates PPG recordings from sixteen public sources, harmonizing annotations into 12 downstream tasks. PulseLM contains over one million PPG segments paired with nearly 2.5 million question-answer pairs, aiming to standardize physiological inference and multimodal model benchmarking. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Establishes a new foundation for multimodal models that integrate physiological data with language understanding.

RANK_REASON This is a research paper introducing a new dataset and benchmark for PPG-text learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• PulseLM
• Manh Pham Hung
• arXiv

• paper
• other