Search arXiv⌕ Search

arXiv subjects

Yan Wu

Publications and source records attributed to Yan Wu.

151 records · Page 9Linked to original sources
arXiv2014

Contractive De-noising Auto-encoder

Auto-encoder is a special kind of neural network based on reconstruction. De-noising auto-encoder (DAE) is an improved auto-encoder which is robust to the input by corrupting the original data first and then reconstructing the original input by minimizing the reconstruction error function. And contractive auto-encoder (CAE) is another kind of improved auto-encoder to learn robust feature by introducing the Frobenius norm of the Jacobean matrix of the learned feature with respect to the original input. In this paper, we combine de-noising auto-encoder and contractive auto- encoder, and propose another improved auto-encoder, contractive de-noising auto- encoder (CDAE), which is robust to both the original input and the learned feature. We stack CDAE to extract more abstract features and apply SVM for classification. The experiment result on benchmark dataset MNIST shows that our proposed CDAE performed better than both DAE and CAE, proving the effective of our method.

cs.LG↗
arXiv2014

Some Comments on the Holographic Heavy Quark Potential in a Thermal Bath

The heavy quark potential of a thermal Yang-Mills theory in strong coupling limit is explored in terms of the holographic principle. With a fairly general AdS/QCD metric the heavy quark potential displays a kink-like screening in the plasma phase. This behavior may conflict the causality of a field theory that is mathematically equivalent to the thermal Yang-Mills.

hep-ph↗
arXiv2013

Spectral Classification Using Restricted Boltzmann Machine

In this study, a novel machine learning algorithm, restricted Boltzmann machine (RBM), is introduced. The algorithm is applied for the spectral classification in astronomy. RBM is a bipartite generative graphical model with two separate layers (one visible layer and one hidden layer), which can extract higher level features to represent the original data. Despite generative, RBM can be used for classification when modified with a free energy and a soft-max function. Before spectral classification, the original data is binarized according to some rule. Then we resort to the binary RBM to classify cataclysmic variables (CVs) and non-CVs (one half of all the given data for training and the other half for testing). The experiment result shows state-of-the-art accuracy of 100%, which indicates the efficiency of the binary RBM algorithm.

cs.LG↗
arXiv2008

Comparative Study of Open-loop Transmit Diversity Schemes for Four Transmit Antennas in Coded OFDM Systems

We compare four open-loop transmit diversity schemes in a coded Orthogonal Frequency Division Multiplexing (OFDM) system with four transmit antennas, namely cyclic delay diversity (CDD), Space-Time Block Code (STBC, Alamouti code is used) with CDD, Quasi-Orthogonal STBC (QO-STBC) and Minimum-Decoding-Complexity QOSTBC (MDC-QOSTBC). We show that in a coded system with low code rate, a scheme with spatial transmit diversity of second order can achieve similar performance to that with spatial transmit diversity of fourth order due to the additional diversity provided by the phase shift diversity with channel coding. In addition, we also compare the decoding complexity and other features of the above four mentioned schemes, such as the requirement for the training signals, hybrid automatic retransmission request (HARQ), etc. The discussions in this paper can be readily applied to future wireless communication systems, such as mobile systems beyond 3G, IEEE 802.11 wireless LAN, or IEEE 802.16 WiMAX, that employ more than two transmit antennas and OFDM.

cs.IT↗
arXiv2006

Envelope Expansion with Core Collapse. III. Similarity Isothermal Shocks in a Magnetofluid

We explore MHD solutions for envelope expansions with core collapse (EECC) with isothermal MHD shocks in a quasi-spherical symmetry and outline potential astrophysical applications of such magnetized shock flows. MHD shock solutions are classified into three classes according to the downstream characteristics near the core. Class I solutions are those characterized by free-fall collapses towards the core downstream of an MHD shock, while Class II solutions are those characterized by Larson-Penston (LP) type near the core downstream of an MHD shock. Class III solutions are novel, sharing both features of Class I and II solutions with the presence of a sufficiently strong magnetic field as a prerequisite. Various MHD processes may occur within the regime of these isothermal MHD shock similarity solutions, such as sub-magnetosonic oscillations, free-fall core collapses, radial contractions and expansions. We can also construct families of twin MHD shock solutions as well as an `isothermal MHD shock' separating two magnetofluid regions of two different yet constant temperatures. The versatile behaviours of such MHD shock solutions may be utilized to model a wide range of astrophysical problems, including star formation in magnetized molecular clouds, MHD link between the asymptotic giant branch phase to the proto-planetary nebula phase with a hot central magnetized white dwarf, relativistic MHD pulsar winds in supernova remnants, radio afterglows of soft gamma-ray repeaters and so forth.

astro-ph↗
arXiv2003

Algorithm and performance of a clinical IMRT beam-angle optimization system

This paper describes the algorithm and examines the performance of an IMRT beam-angle optimization (BAO) system. In this algorithm successive sets of beam angles are selected from a set of predefined directions using a fast simulated annealing (FSA) algorithm. An IMRT beam-profile optimization is performed on each generated set of beams. The IMRT optimization is accelerated by using a fast dose calculation method that utilizes a precomputed dose kernel. A compact kernel is constructed for each of the predefined beams prior to starting the FSA algorithm. The IMRT optimizations during the BAO are then performed using these kernels in a fast dose calculation engine. This technique allows the IMRT optimization to be performed more than two orders of magnitude faster than a similar optimization that uses a convolution dose calculation engine.

physics.med-ph↗