Deep Learning
IPMix: Label-Preserving Data Augmentation Method for Training Robust Classifiers
Data augmentation has been proven effective for training high-accuracy convolu-tional neural network classifiers by preventing overfitting.
Gaussian Process Probes (GPP) for Uncertainty-Aware Probing
Understanding which concepts models can and cannot represent has been fundamental to many tasks: from effective and responsible use of models to detecting out of distribution data.
Riemannian Residual Neural Networks
Recent methods in geometric deep learning have introduced various neural networks to operate over data that lie on Riemannian manifolds.
Wasserstein Quantum Monte Carlo: A Novel Approach for Solving the Quantum Many-Body Schrรถdinger Equation
More specifically, we propose "Wasserstein Quantum Monte Carlo"
Understanding Transformer Predictions Through Memory Efficient Attention Manipulation
Most crucially, they require prohibitively large amounts of additional memory since they rely on backpropagation which allocates almost twice as much GPU memory as the forward pass. This renders it difficult, if not impossible, to use explanations in production.
SEENN: Towards Temporal Spiking Early-Exit Neural Networks Yuhang Li Yale University New Haven, CT, USA
Spiking Neural Networks (SNNs) have recently become more popular as a biologically plausible substitute for traditional Artificial Neural Networks (ANNs).
