CRITS is an intrinsically interpretable time series classifier whose local saliency maps are the exact per-sample weights of the model, obtained without gradients, perturbations, or upsampling.
IEEE Transactions on Artificial Intel- ligence 1(1), 47–61 (2020)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
CRITS: Convolutional Rectifier for Interpretable Time Series Classification
CRITS is an intrinsically interpretable time series classifier whose local saliency maps are the exact per-sample weights of the model, obtained without gradients, perturbations, or upsampling.