Search papers, labs, and topics across Lattice.
Southeast University
3
0
7
Suppressing background noise with heatmap-guided positional embeddings slashes transformer detector parameters by 59% without sacrificing accuracy in small object detection.
You don't need massive models to find tiny cracks: a signal-aware lightweight architecture can outperform heavier detectors in identifying faint subsurface defects in GPR data.
By actively exploring knowledge graphs with a differentiable neural-symbolic approach, NeuroSymActive achieves strong KGQA accuracy while drastically reducing the computational cost of graph lookups.
