Brief

Researchers have developed EMMA, a novel physics-informed framework capable of extracting multiple physical parameters from multimodal data, including video, audio, and chart-based time-series observations. This approach overcomes limitations of prior video-only methods by jointly inferring explicit parameters, latent dynamics, and calibration invariants within a continuous-time model. EMMA utilizes a Liquid Time-Constant network and a physics-constrained loss to ensure consistency with governing differential equations, demonstrating robust multi-parameter recovery across diverse scenarios and outperforming existing baselines. AI

Researchers have developed EMMA, a novel framework capable of extracting multiple physical parameters from raw video, audio, and image data. This physics-informed system utilizes a Liquid Time-Constant network to learn latent dynamics while enforcing consistency with governing differential equations. EMMA demonstrates robust multi-parameter recovery across various benchmarks and real-world systems, outperforming existing single-modality and equation-discovery methods. AI