Search papers, labs, and topics across Lattice.
3
0
4
0
Jointly leveraging selection-based and reinforcement-learning-based alignment can transform medical image captioning into a more trustworthy diagnostic tool.
Clinically structured rewards can enhance medical image captioning accuracy, yielding up to 5.8% better factuality in generated descriptions.
Continuity-driven regularization can boost industrial defect detection performance by over 21 percentage points with minimal training data.
