Statistical Learning
Regularized EM Algorithms: A Unified Framework and Statistical Guarantees
Xinyang Yi, Constantine Caramanis
Neural Information Processing Systems http://nips.cc/
Tree-Sliced Variants of Wasserstein Distances
Tam Le, Makoto Yamada, Kenji Fukumizu, Marco Cuturi
Neural Information Processing Systems http://nips.cc/
. The TSW kernel is
Although Prop. 2 follows from Prop. 1, it follows the idea An upper bound on the Euclidean OT[...] The We will insist more on the importance of sampling tree metrics randomly, both for low-dimensional in 6.1 Definite-negativity is mentioned and highlighted[...] explain why is it important Is this to ensure that the kernel is positive-definite? This is why kernel methods kick in from .6 (or Gaussian processes as per Reviewer #2's suggestion). Indeed, averaging of negative definite functions is trivially negative definite. We used the farthest-point clustering due to its fast computation, i.e.
Lifelong Learning with Non-i.i.d. Tasks
Anastasia Pentina, Christoph H. Lampert
Neural Information Processing Systems http://nips.cc/
Breaking the Communication-Privacy-Accuracy Trilemma
When data is distributed across multiple devices, communication cost often becomes a bottleneck of modern machine learning tasks [38]. This is even more so in federated learning type settings, where communication occurs over bandwidth-limited wireless links [31].
Combinatorial Bayesian Optimization using the Graph Cartesian Product Changyong Oh1 Jakub M. T omczak
Applying BO with GPs on combinatorial spaces is, therefore, not straightforward.
Fast Distributed k-Center Clustering with Outliers on Massive Data
Gustavo Malkomes, Matt J. Kusner, Wenlin Chen, Kilian Q. Weinberger, Benjamin Moseley
Neural Information Processing Systems http://nips.cc/
Regularization-Free Estimation in Trace Regression with Symmetric Positive Semidefinite Matrices
Martin Slawski, Ping Li, Matthias Hein
Trace regression models have received considerable attent ion in the context of matrix completion, quantum state tomography, and compress ed sensing. Estimation of the underlying matrix from regularization-based approaches promoting low-rankedness, notably nuclear norm regularization, hav e enjoyed great popularity. In this paper, we argue that such regularization may no l onger be necessary if the underlying matrix is symmetric positive semidefinite ( spd) and the design satisfies certain conditions. In this situation, simple lea st squares estimation subject to an spd constraint may perform as well as regularization-based app roaches with a proper choice of regularization parameter, which ent ails knowledge of the noise level and/or tuning. By contrast, constrained least s quares estimation comes without any tuning parameter and may hence be preferred due t o its simplicity.