We study a mini-batch diversification scheme for stochastic gradient descent (SGD). While classical SGD relies on uniformly sampling data points to form a mini-batch, we propose a non-uniform sampling scheme based on the Determinantal Point Process (DPP). The DPP relies on a similarity measure between data points and gives low probabilities to mini-batches which contain redundant data, and high...

OBJECTIVES Findings natural products with antioxidant and antiapoptotic properties has been one of the interesting challenges in the search for the treatment of neurodegenerative diseases including ischemic stroke. Serum/glucose deprivation (SGD) has been used as a model for the understanding of the molecular mechanisms of neuronal damage during ischemia in vitro and for the expansion of neurop...

We study a mini-batch diversification scheme for stochastic gradient descent (SGD). While classical SGD relies on uniformly sampling data points to form a mini-batch, we propose a non-uniform sampling scheme based on the Determinantal Point Process (DPP). The DPP relies on a similarity measure between data points and gives low probabilities to mini-batches which contain redundant data, and high...

Matrix factorization (MF) has been widely used in e.g., recommender systems, topic modeling and word embedding. Stochastic gradient descent (SGD) is popular in solving MF problems because it can deal with large data sets and is easy to do incremental learning. We observed that SGD for MF is memory bound. Meanwhile, single-node CPU systems with caching performs well only for small data sets; dis...

Coastal hydrogeologists and oceanographers now recognize the potentially significant contribution that submarine groundwater discharge (SGD) could make to the coastal ocean. SGD may be both volumetrically and chemically important to coastal water and chemical budgets. A worldwide compilation of observed SGD shows that groundwater seepage from the land to the ocean occurs in many environments al...

In this paper, we propose a simple variant of the original stochastic variance reduction gradient (SVRG) [1], where hereafter we refer to as the variance reduced stochastic gradient descent (VR-SGD). Different from the choices of the snapshot point and starting point in SVRG and its proximal variant, Prox-SVRG [2], the two vectors of each epoch in VRSGD are set to the average and last iterate o...

The fate of nitrogen entering Long Island Sound Embayments is dependent on both volume of submarine groundwater discharge (SGD) and geochemical transformations in the coastal aquifer. Previous work in Stony Brook Harbor revealed the importance of sediment composition in the subtidal and offshore areas of the harbor; SGD rate in the subtidal is higher than offshore areas, but experiences only 30...

Objective(s): Rutin is a flavonoid with potent antioxidant property, which exhibited cytoprotective effects in several models of neuronal injury. This work aimed to examine whether rutin can protect neurons against oxidative DNA damage caused by serum/glucose deprivation (SGD) as an in vitro model of neurodegeneration and ischemia. Materials and Metho...

Background: One of the most important issues is the best method to teach embryology course to medical students. Small group discussion (SGD) were used to working together, integral to learningdeveloping intellectual skills and interactive learning experience. Methods: The 72 medical students were equally randomizedto the SGD (group I) and usual lecture based teaching (LBT), (group II) in genera...