Search papers, labs, and topics across Lattice.
1
0
A Kolmogorov-Arnold Network with learnable B-spline activations outperforms a Multi-Layer Perceptron by 44% in cumulative drift error while using 6.9 times fewer parameters.
