We study the properties of common loss surfaces through their Hessian matrix. In particular, in the context of deep learning, we empirically show that the spectrum of the Hessian is composed of two parts: (1) the bulk centered near zero, (2) and outliers away from the bulk. We present numerical evidence and mathematical justifications to the following conjectures laid out by Sagun et al. (2016)...

Convolutional neural networks (CNNs) are powerful tools for classification of visual inputs. An important property of CNN is its restriction to local connections and sharing of local weights among different locations. In this paper, we consider the definition of appropriate local neighborhoods in CNN.We provide a theoretical analysis that justifies the traditional square filter used in CNN for ...

In this paper, we address the problem of cost-efficient inference for non-linear operations in deep neural networks (DNNs), in particular, the exponential function e in softmax layer of DNNs for object detection. The goal is to minimize the hardware cost in terms of energy and area, while maintaining the application accuracy. To this end, we introduce Piecewise Linear Function (PLF) for approxi...

Background & Aims of the Study:  Corrosion and scaling are the most common problems of water facilities. Corrosion imposes many financial and health loss on system and consumer respectively, scaling also causes clogging and head loss in network. The purpose of this study is to determine the corrosion and scaling potential of water supply sources of Qom villages. Materials &...

background: neisseria meningitidis, a life-threatening human pathogen with the potential to cause large epidemics, can be isolated from the nasopharynx of 5–15% of adults. the aim of the current study was to evaluate biophysical and biochemical properties and immunological aspects of chimeric acyl-carrier protein-macrophage infectivity potentiator protein-type iv pilus biogenesis protein antige...

BACKGROUND Education and policy measures within schools are valuable strategies to promote health. This study explored views of experienced food-related educators, researchers and policy-makers regarding their recommended strategies to improve Australian adolescents' knowledge of nutrition and food systems (N&FS). METHODS Semi-structured interviews were conducted with twenty-one experienc...

Fructose-1,6-bisphosphatase, a key enzyme in gluconeogenesis, is subject to metabolic regulation. The human muscle isozyme is significantly more sensitive towards the allosteric inhibitor, AMP, than the liver isoform. Here we report crystal structures and kinetic studies for wild-type human muscle Fru-1,6-Pase, the AMP-bound (1.6 Å), and product-bound complexes of the Q32R mutant, which was fir...

Adaptive Model Predictive Transmission Control Protocol (AMP-TCP) as a new TCP delay-based congestion control algorithm is introduced. Both aspects of design and implementation of the algorithm are described using simulations on the ns-2 network simulator. The design stage is composed of two steps. First, a recursive system identification approach is proposed to capture the network delay dynami...

In the present study, a reusable graphene oxide (GO) based dispersive-solid phase extraction (d-SPE) was synthesized and applied for the analysis of trace ampicillin sodium (AMP) and clindamycin hydrochloride (CLI) in water samples followed by high performance liquid chromatography-UV detection (HPLC-UV). Batch experiments were conducted to investigate the effects of pH and volume of the sample...