Tag: contrastive learning

An international research team has developed CLSS, a protein language model that unites amino acid sequence and structural information into ...

Swansea University researchers have developed a hierarchical contrastive learning framework that combines scarce GP-recorded labels with abundant rule-based eGFR labels ...

A new dual-view graph neural network called HD-SKRG achieves state-of-the-art few-shot molecular property prediction by combining hierarchical atom and functional-group ...

Researchers at Brown University have developed SCIGMA, a deep learning framework that integrates up to five spatial omics modalities across ...

Researchers have unveiled SyncDIT, a diffusion-transformer framework that generates high-fidelity video synchronized frame-by-frame to complex audio such as musical performances, ...

Researchers have developed DKF-DST, a two-stage model that uses contrastive slot selection and dynamic knowledge fusion to significantly improve multi-domain ...

Researchers have developed a deep learning framework that generates clinically faithful diagnostic reports from low-resolution SPECT bone scans by combining ...

A hybrid AI system forecast preictal EEG activity in newborns with high recall while identifying the brain regions influencing its ...