Statistical Learning
SMYRF: Efficient Attention using Asymmetric Clustering
We propose a novel type of balanced clustering algorithm to approximate attention.
federated_pca_paper_neurips.pdf
Despite of this, federated adaptations of classical data analysis techniques are still largely missing.
Stein Variational Gradient Descent with Matrix-Valued Kernels
On the other hand, standard SVGD only uses the first order gradient information, and can not leverage the advantage of the second order methods, such as Newton's method and natural gradient, to achieve better performance on challenging problems with complex loss landscapes or domains.
Certainty Equivalence is Efficient for Linear Quadratic Control
Horia Mania, Stephen Tu, Benjamin Recht
Neural Information Processing Systems http://nips.cc/
main
Learning-based methodologies increasingly find applications in safety-critical domains like autonomous driving and medical robotics.
Adversarial Training and Robustness for Multiple Perturbations
For other perturbations, these defenses offer no guarantees and, at times, even increase the model's vulnerability.
Copula Multi-label Learning
Unfortunately, the statistical properties of existing multi-label dependency modelings are still not well understood.
Residual Flows for Invertible Generative Modeling
Tian Qi Chen, Jens Behrmann, David K. Duvenaud, Joern-Henrik Jacobsen
We give a tractable unbiased estimate of the log density using a "Russian roulette" estimator, and reduce the memory required
Implicit Regularization for Optimal Sparse Recovery
Tomas Vaskevicius, Varun Kanade, Patrick Rebeschini
Hence the total running cost is O(nd), which is the cost to store/read the data in/from memory. These results attest that there are regimes where optimal methods for sparse linear regression exist.