Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Corresponding author, VNU University of Engineering and Technology
3
0
5
CANOLA achieves up to 52% error reduction in label correction, enabling even simple models to outperform complex approaches by 67%.
Relabeler achieves a remarkable 58% improvement in label correction precision, transforming how we handle noisy labels in machine learning.
Key contribution not extracted.