wave arrival time

New Machine Learning Approach Handles Noisy Labels in Seismic Data, Boosts Accuracy by 28.8%

Researchers introduced LaNCoR, a machine learning method that effectively handles noisy labels in seismic signal processing by aligning feature and label representations. Tested on real microseismic data, it improved performance by up to 28.8%, offering a cost-effective alternative to large-scale training sets.