SHAP Outperforms LIME in Explaining Industrial Image Quality Control Models
Researchers evaluated LIME and SHAP interpretability methods on a deep learning model for industrial watch stand inspection, finding that SHAP ...
Researchers evaluated LIME and SHAP interpretability methods on a deep learning model for industrial watch stand inspection, finding that SHAP ...
A neural-network surrogate workflow accelerated four empirical geophysical models while achieving sub-percent mean approximation errors in tested Earth and space ...
Researchers have introduced Global Weight Deep Equilibrium Attention to enhance the predictive accuracy of neural network surrogates for mesh-based numerical ...
A reliability-based optimization study reveals that retaining wall cost and embodied carbon align closely while surrogate models unexpectedly slow down ...
A new probabilistic framework using average maximum failure probability quantifies cumulative slope failure risk under stochastic rainfall over specified exposure ...
A new graph clustering methodology enables flexible hybrid simulation of district heating networks, balancing computational speed with physical accuracy through ...
A new framework called PoAC allows users to define specific clustering goals, outperforming traditional AutoML methods in visualization and anomaly ...
Researchers have shown that randomly shuffling embedded patch tokens inside vision transformer surrogate models during adversarial example generation substantially improves ...
A new editorial collection documents how artificial intelligence is reshaping particle accelerator science, from surrogate models already in practical use ...
Researchers built regressor chain surrogate models that predict pressure and stress fields across varied wing planforms with modest prediction errors ...
© 2026 Scienmag - Science Magazine
© 2026 Scienmag - Science Magazine