The success of deep learning has led to a rising interest in the generalization property of the stochastic gradient descent (SGD) method, and stability is one popular approach to study it. Existing works based on stability have studied nonconvex loss functions, but only considered the generalization error of the SGD in expectation. In this paper, we establish various generalization error bounds...

Allan Dovigi: HIV SALIVARY GLAND DISEASE: A ROLE FOR VIRAL INFECTION (Under the Direction of Dr. Webster Cyriaque) HIV-associated salivary gland disease (HIV SGD) is an AIDS defining condition associated with significant morbidity and lymphoma development in HIV-positive individuals. Understanding HIV SGD becomes increasingly important as the burden of HIV disease expands globally. The epidemio...

background the forced oscillation technique (fot) is a non-invasive means of measuring respiratory resistance and reactance. we tested our hypothesis that endotracheal intubation would cause more substantial preoperative increases in fot parameters than a supraglottic airway device (sgd). methods forty patients requiring general anesthesia and mechanical ventilation for transurethral bladder tu...

Background. Shaoyao-Gancao Decoction (SGD), a well-known traditional Chinese medicine prescription, has been widely used to treat adenomyosis, dysmenorrhea, abdominal pain, and inflammation in Asia. However, the mechanism underlying the effectiveness of SGD in the treatment of adenomyosis still remains elusive. The present study aimed to investigate the bioactivity of SGD and its underlying mol...

[1] Magnitudes of terrestrial (fresh) and marine (saline) sources of submarine groundwater discharge (SGD) are estimated for a transect across Indian River Lagoon, Florida. Two independent techniques (seepage meters and pore water Cl concentrations) show terrestrial SGD decreases linearly to around 22 m offshore, and these techniques, together with a model based on the width of the outflow face...

Abstract Stochastic Gradient Descent (SGD) is widely used in machine learning problems to efficiently perform empirical risk minimization, yet, in practice, SGD is known to stall before reaching the actual minimizer of the empirical risk. SGD stalling has often been attributed to its sensitivity to the conditioning of the problem; however, as we demonstrate, SGD will stall even when applied to ...

We revisit the choice of SGD for training deep neural networks by reconsidering the appropriate geometry in which to optimize the weights. We argue for a geometry invariant to rescaling of weights that does not affect the output of the network, and suggest Path-SGD, which is an approximate steepest descent method with respect to a path-wise regularizer related to max-norm regularization. Path-S...

Submarine groundwater discharge (SGD) into a shallow lagoon on the west coast of Mauritius Island (Flic-en-Flac) was investigated using radioactive ((3)H, (222)Rn, (223)Ra, (224)Ra, (226)Ra, (228)Ra) and stable ((2)H, (18)O) isotopes and nutrients. SGD intercomparison exercises were carried out to validate the various approaches used to measure SGD including radium and radon measurements, seepa...

Submarine groundwater discharge (SGD) is neither well understood nor commonly investigated in Hawai‘i, but it is recognized as a potential pollution source to coastal environments. Between 1998 and 2000, this study located and quantified both total SGD and the terrestrial SGD fraction ( ftgw) in Kahana Bay, O‘ahu. CTD casts were used to profile the water structure and identify potential areas o...

In this paper, we propose a simple variant of the original SVRG, called variance reduced stochastic gradient descent (VR-SGD). Unlike the choices of snapshot and starting points in SVRG and its proximal variant, Prox-SVRG, the two vectors of VR-SGD are set to the average and last iterate of the previous epoch, respectively. The settings allow us to use much larger learning rates, and also make ...