ABSTRACT
“A CROSS MODAL MULTI ATTENTION MECHANISM BASED SYSTEM AND METHOD FOR PROTEIN FUNCTION PREDICTION”
The present invention relates to a cross modal multi attention mechanism based system (100) for protein function prediction, in which the input module (1) receives a protein sequence data, a protein structure data and a protein-protein interaction network data of a protein sample from an operator, the plurality of data processing module include a feature vector generator module (2), a feature extraction module (3), a protein-protein interaction feature set module (4), cross modal multi attention module (5) and a neural network based hierarchical classification module (6) for classifying a function of said protein sample by processing said relative attention distribution value while deploying much fewer trainable parameters than the other methods.
Figure 1 on sheet no. 1 of the drawings may accompany the abstract when published.
