arXiv subjects

Hao Wang

Publications and source records attributed to Hao Wang.

At least 1,603 records · Page 89Linked to original sources

arXiv2019

The MBPEP: a deep ensemble pruning algorithm providing high quality uncertainty prediction

Machine learning algorithms have been effectively applied into various real world tasks. However, it is difficult to provide high-quality machine learning solutions to accommodate an unknown distribution of input datasets; this difficulty is called the uncertainty prediction problems. In this paper, a margin-based Pareto deep ensemble pruning (MBPEP) model is proposed. It achieves the high-quality uncertainty estimation with a small value of the prediction interval width (MPIW) and a high confidence of prediction interval coverage probability (PICP) by using deep ensemble networks. In addition to these networks, unique loss functions are proposed, and these functions make the sub-learners available for standard gradient descent learning. Furthermore, the margin criterion fine-tuning-based Pareto pruning method is introduced to optimize the ensembles. Several experiments including predicting uncertainties of classification and regression are conducted to analyze the performance of MBPEP. The experimental results show that MBPEP achieves a small interval width and a low learning error with an optimal number of ensembles. For the real-world problems, MBPEP performs well on input datasets with unknown distributions datasets incomings and improves learning performance on a multi task problem when compared to that of each single model.

cs.LG↗

arXiv2019

Bidirectional Inference Networks: A Class of Deep Bayesian Networks for Health Profiling

We consider the problem of inferring the values of an arbitrary set of variables (e.g., risk of diseases) given other observed variables (e.g., symptoms and diagnosed diseases) and high-dimensional signals (e.g., MRI images or EEG). This is a common problem in healthcare since variables of interest often differ for different patients. Existing methods including Bayesian networks and structured prediction either do not incorporate high-dimensional signals or fail to model conditional dependencies among variables. To address these issues, we propose bidirectional inference networks (BIN), which stich together multiple probabilistic neural networks, each modeling a conditional dependency. Predictions are then made via iteratively updating variables using backpropagation (BP) to maximize corresponding posterior probability. Furthermore, we extend BIN to composite BIN (CBIN), which involves the iterative prediction process in the training stage and improves both accuracy and computational efficiency by adaptively smoothing the optimization landscape. Experiments on synthetic and real-world datasets (a sleep study and a dermatology dataset) show that CBIN is a single model that can achieve state-of-the-art performance and obtain better accuracy in most inference tasks than multiple models each specifically trained for a different task.

stat.ML↗

arXiv2019

A Particle-Hole-Symmetric Model for a Paired Fractional Quantum Hall State in a Half-filled Landau Level

The fractional quantum Hall effect (FQHE) observed at half filling of the second Landau level is believed to be caused by a pairing of composite fermions captured by the Moore-Read Pfaffian wave function. The generating Hamiltonian for the Moore-Read Pfaffian is a purely three-body model that breaks particle-hole symmetry and lacks other properties, such as dominate two-body repulsive interactions, expected from a physical model of the FQHE. We use exact diagonalization to study the low energy states of a more physical two-body generator model derived from the three-body model. We find that the two-body model exhibits the essential features expected from the Moore-Read Pfaffian: pairing, non-Abelian anyon excitations, and a neutral fermion mode. The model also satisfies constraints expected for a physical model of the FQHE at half-filling because it is: short range, spatially decaying, particle-hole symmetric, and supports a roton mode with a robust spectral gap in the thermodynamic limit. Hence, this two-body model offers a bridge between artificial three-body generator models for paired states and the physical Coulomb interaction and can be used to further explore properties of non-Abelian physics in the FQHE.

cond-mat.str-el↗

arXiv2019

A general model for dynamic contact angle over full speed regime

The prevailing models for advancing dynamic contact angle are under intensive debates, and the fitting performances are far from satisfying in practice. The present study proposes a model based on the recent understanding of the multi-scale structure and the local friction at the contact line. The model has unprecedented fitting performance for dynamic contact angle over wide spreading speed regime across more than five orders of magnitude. The model also well applies for non-monotonous angle variations which have long been considered as abnormal. The model has three fitting parameters, and two of them are nearly predictable at this stage. The one denoting the multi-scale ratio was nearly constant and the one representing the primary frictional coefficient, has a simple correlation with the liquid bulk viscosity.

cond-mat.soft↗

arXiv2019

The origin of the vanadium dioxide transition entropy

The reversible metal-insulator transition in VO$_2$ at $T_\text{C} \approx 340$ K has been closely scrutinized yet its thermodynamic origin remains ambiguous. We discuss the origin of the transition entropy by calculating the electron and phonon contributions at $T_\text{C}$ using density functional theory. The vibration frequencies are obtained from harmonic phonon calculations, with the soft modes that are imaginary at zero temperature renormalized to real values at $T_\text{C}$ using experimental information from diffuse x-ray scattering at high-symmetry wavevectors. Gaussian Process Regression is used to infer the transformed frequencies for wavevectors across the whole Brillouin zone, and in turn compute the finite temperature phonon partition function to predict transition thermodynamics. Using this method, we predict the phase transition in VO$_2$ is driven five to one by phonon entropy over electronic entropy, and predict a total transition entropy that accounts for $95$ % of the calorimetric value.

cond-mat.mtrl-sci↗

arXiv2019

A reduction-to-absurdity approach using absolutely smooth solid surfaces to unveil the origins of wetting

Contact angle hysteresis and generation of dynamics angle are two fundamental phenomena about the contact angle deviation from the equilibrium state. Roughness on the solid surface, disjoining pressure in the thin film, and liquid-solid adhesion have all been considered as the origins of the phenomena. This work for the first time made a reduction to absurdity by employing absolutely smooth solid surfaces in ultra-large-scale molecular dynamic simulation. The results showed that the equilibrium angles were well established on the absolutely smooth surface just as regular, while the hysteresis and the dynamic angle vanished. The critical structure of the convex nanobending for advancing contact lines vanished as well. In contrast, the solids that made of atoms, even at the minimum roughness, would bring significant angle deviation and convex nanobending. A 3D observation was further made using state-of-the-art helium ion microscopy for the first time revealing the ubiquitous nanoscopic distortion along the contact line on atomically smooth surfaces. The results answer the question of the origin of the angle deviation, that the hysteresis and dynamic angle can be unified to originate from the friction, either static or dynamic; and at the same time resolve the puzzle of the hysteresis existence on reported smooth surfaces, showing that even the minimum i.e. atomic roughness is great enough to be effective on the contact line. The results are against the disjoining pressure theory which ignored the roughness and predicted the occurrence of the hysteresis on absolutely smooth solids.

cond-mat.soft↗

arXiv2019

A Hardware Friendly Unsupervised Memristive Neural Network with Weight Sharing Mechanism

Memristive neural networks (MNNs), which use memristors as neurons or synapses, have become a hot research topic recently. However, most memristors are not compatible with mainstream integrated circuit technology and their stabilities in large-scale are not very well so far. In this paper, a hardware friendly MNN circuit is introduced, in which the memristive characteristics are implemented by digital integrated circuit. Through this method, spike timing dependent plasticity (STDP) and unsupervised learning are realized. A weight sharing mechanism is proposed to bridge the gap of network scale and hardware resource. Experiment results show the hardware resource is significantly saved with it, maintaining good recognition accuracy and high speed. Moreover, the tendency of resource increase is slower than the expansion of network scale, which infers our method's potential on large scale neuromorphic network's realization.

cs.ET↗

arXiv2018

Adaptive QM/MM Coupling for Crystalline Defects

QM (quantum mechenics) and MM (molecular mechenics) coupling methods are widely used in simulations of crystalline defects. In this paper, we construct a residual based a posteriori error indicator for QM/MM coupling approximations. We prove the reliability of the error indicator (upper bound of the true approximation error) and develop some sampling techniques for its efficient calculation. Based on the error indicator and Dörfler marking strategy, we design an adaptive QM/MM algorithm for crystalline defects and demonstrate the efficiency with some numerical experiments.

math.NA↗

arXiv2018

Compact Graphene Plasmonic Slot Photodetector on Silicon-on-insulator with High Responsivity

Graphene has extraordinary electro-optic properties and is therefore a promising candidate for monolithic photonic devices such as photodetectors. However, the integration of this atom-thin layer material with bulky photonic components usually results in a weak light-graphene interaction leading to large device lengths limiting electro-optic performance. In contrast, here we demonstrate a plasmonic slot graphene photodetector on silicon-on-insulator platform with high-responsivity given the 5 um-short device length. We observe that the maximum photocurrent, and hence the highest responsivity, scales inversely with the slot gap width. Using a dual-lithography step, we realize 15 nm narrow slots that show a 15-times higher responsivity per unit device-length compared to photonic graphene photodetectors. Furthermore, we reveal that the back-gated electrostatics is overshadowed by channel-doping contributions induced by the contacts of this ultra-short channel graphene photodetector. This leads to quasi charge neutrality, which explains both the previously-unseen offset between the maximum photovoltaic-based photocurrent relative to graphenes Dirac point and the observed non-ambipolar transport. Such micrometer compact and absorption-efficient photodetectors allow for short-carrier pathways in next-generation photonic components, while being an ideal testbed to study short-channel carrier physics in graphene optoelectronics.

physics.app-ph↗

arXiv2018

l0-norm Based Centers Selection for Training Fault Tolerant RBF Networks and Selecting Centers

The aim of this paper is to train an RBF neural network and select centers under concurrent faults. It is well known that fault tolerance is a very attractive property for neural networks. And center selection is an important procedure during the training process of an RBF neural network. In this paper, we devise two novel algorithms to address these two issues simultaneously. Both of them are based on the ADMM framework. In the first method, the minimax concave penalty (MCP) function is introduced to select centers. In the second method, an l0-norm term is directly used, and the hard threshold (HT) is utilized to address the l0-norm term. Under several mild conditions, we can prove that both methods can globally converge to a unique limit point. Simulation results show that, under concurrent fault, the proposed algorithms are superior to many existing methods.

cs.LG↗

arXiv2018

Budget Constrained Bidding by Model-free Reinforcement Learning in Display Advertising

Real-time bidding (RTB) is an important mechanism in online display advertising, where a proper bid for each page view plays an essential role for good marketing results. Budget constrained bidding is a typical scenario in RTB where the advertisers hope to maximize the total value of the winning impressions under a pre-set budget constraint. However, the optimal bidding strategy is hard to be derived due to the complexity and volatility of the auction environment. To address these challenges, in this paper, we formulate budget constrained bidding as a Markov Decision Process and propose a model-free reinforcement learning framework to resolve the optimization problem. Our analysis shows that the immediate reward from environment is misleading under a critical resource constraint. Therefore, we innovate a reward function design methodology for the reinforcement learning problems with constraints. Based on the new reward design, we employ a deep neural network to learn the appropriate reward so that the optimal policy can be learned effectively. Different from the prior model-based work, which suffers from the scalability problem, our framework is easy to be deployed in large-scale industrial applications. The experimental evaluations demonstrate the effectiveness of our framework on large-scale real datasets.

cs.AI↗

arXiv2018

Orthogonal Deep Features Decomposition for Age-Invariant Face Recognition

As facial appearance is subject to significant intra-class variations caused by the aging process over time, age-invariant face recognition (AIFR) remains a major challenge in face recognition community. To reduce the intra-class discrepancy caused by the aging, in this paper we propose a novel approach (namely, Orthogonal Embedding CNNs, or OE-CNNs) to learn the age-invariant deep face features. Specifically, we decompose deep face features into two orthogonal components to represent age-related and identity-related features. As a result, identity-related features that are robust to aging are then used for AIFR. Besides, for complementing the existing cross-age datasets and advancing the research in this field, we construct a brand-new large-scale Cross-Age Face dataset (CAF). Extensive experiments conducted on the three public domain face aging datasets (MORPH Album 2, CACD-VS and FG-NET) have shown the effectiveness of the proposed approach and the value of the constructed CAF dataset on AIFR. Benchmarking our algorithm on one of the most popular general face recognition (GFR) dataset LFW additionally demonstrates the comparable generalization performance on GFR.

cs.CV↗

arXiv2018

Porosity-mediated High-performance Thermoelectric Materials

Whether porosity can effectively improve thermoelectric performance is still an open question. Herein we report that thermoelectric performance can be significantly enhanced by creating porosity in n-type Mg3.225Mn0.025Sb1.5Bi0.49Te0.01, with a ZT of ~0.9 at 323 K and ~1.6 at 723 K, making the average ZT much higher for better performance. The large improvement at room temperature is significant considering that such a ZT value is comparable to the best ZT at this temperature in n-type Bi2Te3. The enhancement was mainly from the improved electrical mobility and multi-scale phonon scattering, particularly from the well-dispersed bismuth nano-precipitates in the porous structure. We further extend this approach to other thermoelectric materials such as half-Heuslers Nb0.56V0.24Ti0.2FeSb and Hf0.25Zr0.75NiSn0.99Sb0.01 and Bi0.5Sb1.5Te3 showing similar improvements, further advancing thermoelectric materials for applications.

cond-mat.mtrl-sci↗

arXiv2018

IOHprofiler: A Benchmarking and Profiling Tool for Iterative Optimization Heuristics

IOHprofiler is a new tool for analyzing and comparing iterative optimization heuristics. Given as input algorithms and problems written in C or Python, it provides as output a statistical evaluation of the algorithms' performance by means of the distribution on the fixed-target running time and the fixed-budget function values. In addition, IOHprofiler also allows to track the evolution of algorithm parameters, making our tool particularly useful for the analysis, comparison, and design of (self-)adaptive algorithms. IOHprofiler is a ready-to-use software. It consists of two parts: an experimental part, which generates the running time data, and a post-processing part, which produces the summarizing comparisons and statistical evaluations. The experimental part is build on the COCO software, which has been adjusted to cope with optimization problems that are formulated as functions $f:\mathcal{S}^n \to \R$ with $\mathcal{S}$ being a discrete alphabet of integers. The post-processing part is our own work. It can be used as a stand-alone tool for the evaluation of running time data of arbitrary benchmark problems. It accepts as input files not only the output files of IOHprofiler, but also original COCO data files. The post-processing tool is designed for an interactive evaluation, allowing the user to chose the ranges and the precision of the displayed data according to his/her needs. IOHprofiler is available on GitHub at \url{https://github.com/IOHprofiler}.

cs.NE↗

arXiv2018

Automatic Configuration of Deep Neural Networks with EGO

Designing the architecture for an artificial neural network is a cumbersome task because of the numerous parameters to configure, including activation functions, layer types, and hyper-parameters. With the large number of parameters for most networks nowadays, it is intractable to find a good configuration for a given task by hand. In this paper an Efficient Global Optimization (EGO) algorithm is adapted to automatically optimize and configure convolutional neural network architectures. A configurable neural network architecture based solely on convolutional layers is proposed for the optimization. Without using any knowledge on the target problem and not using any data augmentation techniques, it is shown that on several image classification tasks this approach is able to find competitive network architectures in terms of prediction accuracy, compared to the best hand-crafted ones in literature. In addition, a very small training budget (200 evaluations and 10 epochs in training) is spent on each optimized architectures in contrast to the usual long training time of hand-crafted networks. Moreover, instead of the standard sequential evaluation in EGO, several candidate architectures are proposed and evaluated in parallel, which saves the execution overheads significantly and leads to an efficient automation for deep neural network design.

cs.LG↗

arXiv2018

Deep Inferential Spatial-Temporal Network for Forecasting Air Pollution Concentrations

Air pollution poses a serious threat to human health as well as economic development around the world. To meet the increasing demand for accurate predictions for air pollutions, we proposed a Deep Inferential Spatial-Temporal Network to deal with the complicated non-linear spatial and temporal correlations. We forecast three air pollutants (i.e., PM2.5, PM10 and O3) of monitoring stations over the next 48 hours, using a hybrid deep learning model consists of inferential predictor (inference for regions without air pollution readings), spatial predictor (capturing spatial correlations using CNN) and temporal predictor (capturing temporal relationship using sequence-to-sequence model with simplified attention mechanism). Our proposed model considers historical air pollution records and historical meteorological data. We evaluate our model on a large-scale dataset containing air pollution records of 35 monitoring stations and grid meteorological data in Beijing, China. Our model outperforms other state-of-art methods in terms of SMAPE and RMSE.

cs.LG↗

arXiv2018

Towards a Theory-Guided Benchmarking Suite for Discrete Black-Box Optimization Heuristics: Profiling $(1+λ)$ EA Variants on OneMax and LeadingOnes

Theoretical and empirical research on evolutionary computation methods complement each other by providing two fundamentally different approaches towards a better understanding of black-box optimization heuristics. In discrete optimization, both streams developed rather independently of each other, but we observe today an increasing interest in reconciling these two sub-branches. In continuous optimization, the COCO (COmparing Continuous Optimisers) benchmarking suite has established itself as an important platform that theoreticians and practitioners use to exchange research ideas and questions. No widely accepted equivalent exists in the research domain of discrete black-box optimization. Marking an important step towards filling this gap, we adjust the COCO software to pseudo-Boolean optimization problems, and obtain from this a benchmarking environment that allows a fine-grained empirical analysis of discrete black-box heuristics. In this documentation we demonstrate how this test bed can be used to profile the performance of evolutionary algorithms. More concretely, we study the optimization behavior of several $(1+λ)$ EA variants on the two benchmark problems OneMax and LeadingOnes. This comparison motivates a refined analysis for the optimization time of the $(1+λ)$ EA on LeadingOnes.

cs.NE↗

arXiv2018

A quantized physical framework for understanding the working mechanism of ion channels

A quantized physical framework, called the five-anchor model, is developed for a general understanding of the working mechanism of ion channels. According to the hypotheses of this model, the following two basic physical principles are assigned to each anchor: the polarity change induced by an electron transition and the mutual repulsion and attraction induced by an electrostatic force. Consequently, many unique phenomena, such as fast and slow inactivation, the stochastic gating pattern and constant conductance of a single ion channel, the difference between electrical and optical stimulation (optogenetics), nerve conduction block and the generation of an action potential, become intrinsic features of this physical model. Moreover, this model also provides a foundation for the probability equation used to calculate the results of electrical stimulation in our previous C-P theory.

q-bio.NC↗