Kalman-Based Stochastic Gradient Method with Stop Condition and Insensitivity to Conditioning
نویسندگان
چکیده
منابع مشابه
Kalman-Based Stochastic Gradient Method with Stop Condition and Insensitivity to Conditioning
Abstract. Proximal and stochastic gradient descent (SGD) methods are believed to efficiently minimize large composite objective functions, but such methods have two algorithmic challenges: (1) a lack of fast or justified stopping conditions, and (2) sensitivity to the problem’s conditioning. Second order SGD methods show promise in solving these problems, but they are (3) marred by the complexi...
متن کاملStochastic gradient method with accelerated stochastic dynamics
In this paper, we propose a novel technique to implement stochastic gradient methods, which are beneficial for learning from large datasets, through accelerated stochastic dynamics. A stochastic gradient method is based on mini-batch learning for reducing the computational cost when the amount of data is large. The stochasticity of the gradient can be mitigated by the injection of Gaussian nois...
متن کاملStochastic generalized gradient method with application to insurance risk management
Recently [9] we analyzed important classes of nonsmooth and nonconvex risk control problems which can not be solved by standard optimization techniques. The aim of this article is to develop computational procedures enabling us to bypass some of the obstacles identified in this paper. We illustrate this by using insurance risk processes with insolvency (stopping time).
متن کاملCPSG-MCMC: Clustering-Based Preprocessing method for Stochastic Gradient MCMC
In recent years, stochastic gradient Markov Chain Monte Carlo (SG-MCMC) methods have been raised to process large-scale dataset by iterative learning from small minibatches. However, the high variance caused by naive subsampling usually slows down the convergence to the desired posterior distribution. In this paper, we propose an effective subsampling strategy to reduce the variance based on a ...
متن کاملa comparison of teachers and supervisors, with respect to teacher efficacy and reflection
supervisors play an undeniable role in training teachers, before starting their professional experience by preparing them, at the initial years of their teaching by checking their work within the proper framework, and later on during their teaching by assessing their progress. but surprisingly, exploring their attributes, professional demands, and qualifications has remained a neglected theme i...
15 صفحه اولذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: SIAM Journal on Optimization
سال: 2016
ISSN: 1052-6234,1095-7189
DOI: 10.1137/15m1048239