Self-Supervised Learning by Estimating Twin Class Distributions

Kavli Affiliate: Feng Wang | First 5 Authors: Feng Wang, Tao Kong, Rufeng Zhang, Huaping Liu, Hang Li | Summary: We present TWIST, a simple and theoretically explainable self-supervised representation learning method by classifying large-scale unlabeled datasets in an end-to-end way. We employ a siamese network terminated by a softmax operation to produce twin class […]


Continue.. Self-Supervised Learning by Estimating Twin Class Distributions

Finding Local Minimax Points via (Stochastic) Cubic-Regularized GDA: Global Convergence and Complexity

Kavli Affiliate: Yi Zhou | First 5 Authors: Ziyi Chen, Qunwei Li, Yi Zhou, , | Summary: Standard gradient descent-ascent (GDA)-type algorithms can only find stationary points in nonconvex minimax optimization, which are far more sub-optimal than local minimax points. In this work, we develop GDA-type algorithms that globally converge to local minimax points in […]


Continue.. Finding Local Minimax Points via (Stochastic) Cubic-Regularized GDA: Global Convergence and Complexity

Structure of Stellar Remnants with Coupling to a Light Scalar

Kavli Affiliate: Albert Stebbins | First 5 Authors: Christina Gao, Albert Stebbins, , , | Summary: In this paper we study how a Yukawa coupling of the Standard Model fermions to a light scalar field effects the stellar structure of cold stellar remnants such as neutron stars. We elucidate the stellar structure phenomenology using a […]


Continue.. Structure of Stellar Remnants with Coupling to a Light Scalar

$S^5$: The Orbital and Chemical Properties of One Dozen Stellar Streams

Kavli Affiliate: Alexander P. Ji | First 5 Authors: Ting S. Li, Alexander P. Ji, Andrew B. Pace, Denis Erkal, Sergey E. Koposov | Summary: We report the kinematic, orbital, and chemical properties of 12 stellar streams with no evident progenitors, using line-of-sight velocities and metallicities from the Southern Stellar Stream Spectroscopic Survey ($S^5$), proper […]


Continue.. $S^5$: The Orbital and Chemical Properties of One Dozen Stellar Streams

A Malmquist-like bias in the inferred areas of diamond caustics and consequences for inferred time delays of gravitationally lensed quasars

Kavli Affiliate: Paul Schechter | Summary:Quasars are quadruply lensed only when they lie within the diamond caustic of a lensing galaxy. This precondition produces a Malmquist-like selection effect in observed populations of quadruply lensed quasars, overestimating the true caustic area. The bias toward high values of the inferred logarithmic area, $ln A_inf$, is proportional to […]


Continue.. A Malmquist-like bias in the inferred areas of diamond caustics and consequences for inferred time delays of gravitationally lensed quasars

A Malmquist-like bias in the inferred areas of diamond caustics and consequences for inferred time delays of gravitationally lensed quasars

Kavli Affiliate: Paul L. Schechter | First 5 Authors: Derek Baldwin, Paul L. Schechter, , , | Summary: Quasars are quadruply lensed only when they lie within the diamond caustic of a lensing galaxy. This precondition produces a Malmquist-like selection effect in observed populations of quadruply lensed quasars, overestimating the true caustic area. The bias […]


Continue.. A Malmquist-like bias in the inferred areas of diamond caustics and consequences for inferred time delays of gravitationally lensed quasars

A Malmquist-like bias in the inferred areas of diamond caustics and the resulting bias in inferred time delays for gravitationally lensed quasars

Kavli Affiliate: Paul L. Schechter | First 5 Authors: Derek Baldwin, Paul L. Schechter, , , | Summary: Quadruply lensed quasars are visible only when the source quasar lies within the diamond caustic of the lensing galaxy. This condition creates a Malmquist-like selection effect in the population of observed quadruply lensed quasars, increasing the true […]


Continue.. A Malmquist-like bias in the inferred areas of diamond caustics and the resulting bias in inferred time delays for gravitationally lensed quasars

A global convergence theory for deep ReLU implicit networks via over-parameterization

Kavli Affiliate: Jia Liu | First 5 Authors: Tianxiang Gao, Hailiang Liu, Jia Liu, Hridesh Rajan, Hongyang Gao | Summary: Implicit deep learning has received increasing attention recently due to the fact that it generalizes the recursive prediction rules of many commonly used neural network architectures. Its prediction rule is provided implicitly based on the […]


Continue.. A global convergence theory for deep ReLU implicit networks via over-parameterization

Graph-Guided Network for Irregularly Sampled Multivariate Time Series

Kavli Affiliate: Xiang Zhang | First 5 Authors: Xiang Zhang, Marko Zeman, Theodoros Tsiligkaridis, Marinka Zitnik, | Summary: In many domains, including healthcare, biology, and climate science, time series are irregularly sampled with varying time intervals between successive readouts and different subsets of variables (sensors) observed at different time points. Here, we introduce RAINDROP, a […]


Continue.. Graph-Guided Network for Irregularly Sampled Multivariate Time Series