Goto

Collaborating Authors

 Asia







96671501524948bc3937b4b30d0e57b9-Paper.pdf

Neural Information Processing Systems

BERT is incapable of processing long texts due to its quadratically increasing memory andtimeconsumption. Themost natural waystoaddress thisproblem, such as slicing the text by a sliding window or simplifying transformers, suffer from insufficient long-range attentions orneed customized CUDAkernels.



StabilityAnalysisandGeneralizationBoundsof AdversarialTraining

Neural Information Processing Systems

In adversarial machine learning, deep neural networks can fit the adversarial examples on the training dataset but have poor generalizationability on the test set.



BernNet: LearningArbitraryGraphSpectralFilters viaBernsteinApproximation

Neural Information Processing Systems

Graph neural networks (GNNs) have received extensive attention from researchers due to their excellent performance on various graph learning tasks such as social analysis [24, 17, 29], drug discovery [12, 25], traffic forecasting [18, 3, 6], recommendation system [38, 32] and computer vision[39,4].