arXiv subjects

Wei Gao

Publications and source records attributed to Wei Gao.

At least 307 records · Page 17Linked to original sources

arXiv2018

On the Resistance of Nearest Neighbor to Random Noisy Labels

Nearest neighbor has always been one of the most appealing non-parametric approaches in machine learning, pattern recognition, computer vision, etc. Previous empirical studies partly shows that nearest neighbor is resistant to noise, yet there is a lack of deep analysis. This work presents the finite-sample and distribution-dependent bounds on the consistency of nearest neighbor in the random noise setting. The theoretical results show that, for asymmetric noises, k-nearest neighbor is robust enough to classify most data correctly, except for a handful of examples, whose labels are totally misled by random noises. For symmetric noises, however, k-nearest neighbor achieves the same consistent rate as that of noise-free setting, which verifies the resistance of k-nearest neighbor to random noisy labels. Motivated by the theoretical analysis, we propose the Robust k-Nearest Neighbor (RkNN) approach to deal with noisy labels. The basic idea is to make unilateral corrections to examples, whose labels are totally misled by random noises, and classify the others directly by utilizing the robustness of k-nearest neighbor. We verify the effectiveness of the proposed algorithm both theoretically and empirically.

cs.LG↗

arXiv2018

Generative Adversarial Networks with Decoder-Encoder Output Noise

In recent years, research on image generation methods has been developing fast. The auto-encoding variational Bayes method (VAEs) was proposed in 2013, which uses variational inference to learn a latent space from the image database and then generates images using the decoder. The generative adversarial networks (GANs) came out as a promising framework, which uses adversarial training to improve the generative ability of the generator. However, the generated pictures by GANs are generally blurry. The deep convolutional generative adversarial networks (DCGANs) were then proposed to leverage the quality of generated images. Since the input noise vectors are randomly sampled from a Gaussian distribution, the generator has to map from a whole normal distribution to the images. This makes DCGANs unable to reflect the inherent structure of the training data. In this paper, we propose a novel deep model, called generative adversarial networks with decoder-encoder output noise (DE-GANs), which takes advantage of both the adversarial training and the variational Bayesain inference to improve the performance of image generation. DE-GANs use a pre-trained decoder-encoder architecture to map the random Gaussian noise vectors to informative ones and pass them to the generator of the adversarial networks. Since the decoder-encoder architecture is trained by the same images as the generators, the output vectors could carry the intrinsic distribution information of the original images. Moreover, the loss function of DE-GANs is different from GANs and DCGANs. A hidden-space loss function is added to the adversarial loss function to enhance the robustness of the model. Extensive empirical results show that DE-GANs can accelerate the convergence of the adversarial training process and improve the quality of the generated images.

cs.CV↗

arXiv2018

Efficient Computational Algorithm for Optimal Continuous Experimental Designs

A simple yet efficient computational algorithm for computing the continuous optimal experimental design for linear models is proposed. An alternative proof the monotonic convergence for $D$-optimal criterion on continuous design spaces are provided. We further show that the proposed algorithm converges to the $D$-optimal design. We also provide an algorithm for the $A$-optimality and conjecture that the algorithm convergence monotonically on continuous design spaces. Different numerical examples are used to demonstrated the usefulness and performance of the proposed algorithms.

stat.CO↗

arXiv2018

Dynamic Actuator Selection and Robust State-Feedback Control of Networked Soft Actuators

The design of robots that are light, soft, powerful is a grand challenge. Since they can easily adapt to dynamic environments, soft robotic systems have the potential of changing the status-quo of bulky robotics. A crucial component of soft robotics is a soft actuator that is activated by external stimuli to generate desired motions. Unfortunately, there is a lack of powerful soft actuators that operate through lightweight power sources. To that end, we recently designed a highly scalable, flexible, biocompatible Electromagnetic Soft Actuator (ESA). With ESAs, artificial muscles can be designed by integrating a network of ESAs. The main research gap addressed in this work is in the absence of system-theoretic understanding of the impact of the realtime control and actuator selection algorithms on the performance of networked soft-body actuators and ESAs. The objective of this paper is to establish a framework that guides the analysis and robust control of networked ESAs. A novel ESA is described, and a configuration of soft actuator matrix to resemble artificial muscle fiber is presented. A mathematical model which depicts the physical network is derived, considering the disturbances due to external forces and linearization errors as an integral part of this model. Then, a robust control and minimal actuator selection problem with logistic constraints and control input bounds is formulated, and tractable computational routines are proposed with numerical case studies.

eess.SY↗

arXiv2018

Unorganized Malicious Attacks Detection

Recommender system has attracted much attention during the past decade. Many attack detection algorithms have been developed for better recommendations, mostly focusing on shilling attacks, where an attack organizer produces a large number of user profiles by the same strategy to promote or demote an item. This work considers a different attack style: unorganized malicious attacks, where attackers individually utilize a small number of user profiles to attack different items without any organizer. This attack style occurs in many real applications, yet relevant study remains open. We first formulate the unorganized malicious attacks detection as a matrix completion problem, and propose the Unorganized Malicious Attacks detection (UMA) approach, a proximal alternating splitting augmented Lagrangian method. We verify, both theoretically and empirically, the effectiveness of our proposed approach.

cs.IR↗

arXiv2017

Influence of Bi addition on the property of Ag-Bi nano-composite coatings

Silver (Ag) coatings have been widely used in many industry areas due to their excellent conductivity. However, wider applications of Ag coatings have been hindered by their poor mechanical properties. In this research, to improve the mechanical performance, Ag-Bi nano-composite coatings were prepared by a novel ionic co-discharge method. A systematic study of the microstructure, mechanical properties, electrical conductivity and antibacterial behavior of the resulting coating was performed. The results indicated that after adding an appropriate amount of Bi containing solution into the Ag plating solution, Ag-Bi nano-particles were in-situ formed and distributed uniformly throughout the coating matrix, resulting in a significant improvement in the mechanical properties. The hardness of Ag-Bi coating was increased by 60% compared to that of the pure Ag coating. The corrosion resistance of Ag-Bi coatings was also enhanced. The outcome of this research may find a broader application in electronics, jewelry, aerospace and other industries.

cond-mat.mtrl-sci↗

arXiv2017

Estimation of Treatment Effects for Heterogeneous Matched Pairs Data with Probit Models

Estimating the effect of medical treatments on subject responses is one of the crucial problems in medical research. Matched-pairs designs are commonly implemented in the field of medical research to eliminate confounding and improve efficiency. In this article, new estimators of treatment effects for heterogeneous matched pairs data are proposed. Asymptotic properties of the proposed estimators are derived. Simulation studies show that the proposed estimators have some advantages over the famous Heckman's estimator and inverse probability weighted (IPW) estimator. We apply the proposed methodologies to a blood lead level data set and an acute leukaemia data set.

stat.ME↗

arXiv2017

$k$-Foldability of Words

We extend results regarding a combinatorial model introduced by Black, Drellich, and Tymoczko (2017+) which generalizes the folding of the RNA molecule in biology. Consider a word on alphabet $\{A_1, \overline{A}_1, \ldots, A_m, \overline{A}_m\}$ in which $\overline{A}_i$ is called the complement of $A_i$. A word $w$ is foldable if can be wrapped around a rooted plane tree $T$, starting at the root and working counterclockwise such that one letter labels each half edge and the two letters labeling the same edge are complements. The tree $T$ is called $w$-valid. We define a bijection between edge-colored plane trees and words folded onto trees. This bijection is used to characterize and enumerate words for which there is only one valid tree. We follow up with a characterization of words for which there exist exactly two valid trees. In addition, we examine the set $\mathcal{R}(n,m)$ consisting of all integers $k$ for which there exists a word of length $2n$ with exactly $k$ valid trees. Black, Drellich, and Tymoczko showed that for the $n$th Catalan number $C_n$, $\{C_n,C_{n-1}\}\subset \mathcal{R}(n,1)$ but $k\not\in\mathcal{R}(n,1)$ for $C_{n-1}<k<C_n$. We describe a superset of $\mathcal{R}(n,1)$ in terms of the Catalan numbers by which we establish more missing intervals. We also prove $\mathcal{R}(n,1)$ contains all non-negative integer less than $n+1$.

math.CO↗

arXiv2017

Sign patterns that require $\mathbb{H}_n$ exist for each $n\geq 4$

The refined inertia of a square real matrix $A$ is the ordered $4$-tuple $(n_+, n_-, n_z, 2n_p)$, where $n_+$ (resp., $n_-$) is the number of eigenvalues of $A$ with positive (resp., negative) real part, $n_z$ is the number of zero eigenvalues of $A$, and $2n_p$ is the number of nonzero pure imaginary eigenvalues of $A$. For $n \geq 3$, the set of refined inertias $\mathbb{H}_n=\{(0, n, 0, 0), (0, n-2, 0, 2), (2, n-2, 0, 0)\}$ is important for the onset of Hopf bifurcation in dynamical systems. We say that an $n\times n$ sign pattern ${\cal A}$ requires $\mathbb{H}_n$ if $\mathbb{H}_n=\{\text{ri}(B) | B \in Q({\cal A})\}$. Bodine et al. conjectured that no $n\times n$ irreducible sign pattern that requires $\mathbb{H}_n$ exists for $n$ sufficiently large, possibly $n\ge 8$. However, for each $n \geq 4$, we identify three $n\times n$ irreducible sign patterns that require $\mathbb{H}_n$, which resolves this conjecture.

math.CO↗

arXiv2017

Intention-Net: Integrating Planning and Deep Learning for Goal-Directed Autonomous Navigation

How can a delivery robot navigate reliably to a destination in a new office building, with minimal prior information? To tackle this challenge, this paper introduces a two-level hierarchical approach, which integrates model-free deep learning and model-based path planning. At the low level, a neural-network motion controller, called the intention-net, is trained end-to-end to provide robust local navigation. The intention-net maps images from a single monocular camera and "intentions" directly to robot controls. At the high level, a path planner uses a crude map, e.g., a 2-D floor plan, to compute a path from the robot's current location to the goal. The planned path provides intentions to the intention-net. Preliminary experiments suggest that the learned motion controller is robust against perceptual uncertainty and by integrating with a path planner, it generalizes effectively to new environments and goals.

cs.AI↗

arXiv2017

Revealing photons' past via quantum twisted double-slit experiment

Are quantum states real? This most fundamental question in quantum mechanics has not yet been satisfactorily resolved, although its realistic interpretation seems to have been rejected by various delayed-choice experiments. Here, to address this long-standing issue, we present a quantum twisted double-slit experiment. By exploiting the subluminal feature of twisted photons, the real nature of a photon during its time in flight is revealed for the first time. We found that photons' arrival times were inconsistent with the states obtained in measurements but agreed with the states during propagation. Our results demonstrate that wavefunctions describe the realistic existence and evolution of quantum entities rather than a pure mathematical abstraction providing a probability list of measurement outcomes. This finding clarifies the long-held misunderstanding of the role of wavefunctions and their collapse in the evolution of quantum entities.

quant-ph↗

arXiv2017

Effects of the optimisation of the margin distribution on generalisation in deep architectures

Despite being so vital to success of Support Vector Machines, the principle of separating margin maximisation is not used in deep learning. We show that minimisation of margin variance and not maximisation of the margin is more suitable for improving generalisation in deep architectures. We propose the Halfway loss function that minimises the Normalised Margin Variance (NMV) at the output of a deep learning models and evaluate its performance against the Softmax Cross-Entropy loss on the MNIST, smallNORB and CIFAR-10 datasets.

cs.LG↗

arXiv2017

A Novel Thiophene-fused Ending Group Enabling an Excellent Small Molecule Acceptor for High-performance Fullerene-free Polymer Solar Cells with 11.8% Efficiency

In this article, we designed and synthesized a novel small molecule acceptor of ITCPTC with thiophene-fused ending group by employing a new active methylene precursor of CPTCN. The ITCPTC based polymer solar cells with PBT1-EH as donor achieved very high PCEs of up to 11.8% with a remarkably enhanced fill factor (FF) of 0.751, a near 20% boost in PCE with respect to the ITIC based control device. These values are among the highest PCEs and FFs for PSCs. In the whole study, we made contrasts with ITIC to understand the reasons of excellent performance of ITCPTC-based PSCs through various measurements, such as GIWAXS and RSoXS. We revealed that the simple modification of ITIC into ITCPTC not only change the material electronic structure, but also mediate the material interactions and crystallization, which contribute together to the excellent performance of ITCPTC based PSCs.

physics.chem-ph↗

arXiv2017

Generating strong cylindrical vector pulses via stimulated Brillouin amplification

Light with transverse polarization structure, such as radial and azimuthal polarization, enables and revives lots of applications based on light-matter interaction due to their unique focal properties. To date, studies referring to this topic mainly concentrate in weak-light domain, yet it should have gained more attention in strong-light domain. A main factor contributing to the current situation is that the generation devices cannot afford high power. Here, in this proof-of-principle work, we demonstrate the generation of strong single-longitudinal-mode (SLM) cylindrical vector (CV) short pulses via stimulated Brillouin amplification, specifically, the energy is transferred from a 700 ps pump to a 300 ps Stokes pulse via parametrically generated coherent phonons. After amplification, a 100 mJ-level SLM CV pulse light with 300 ps duration is obtained. Meanwhile, the phase and polarization structures are high fidelity maintained. This result provides a practicable way to generate strong CV light, and by extending this mechanism into multi-pump beam combing system, even an ultra-high intensity CV light can be expected.

physics.optics↗

arXiv2016

Bounds of Zagreb indices and hyper Zagreb indices

The hyper Zagreb index is a kind of extensions of Zagreb index, used for predicting physicochemical properties of organic compounds. Given a graph $G= (V(G), E(G))$, the first hyper-Zagreb index is the sum of the square of edge degree over edge set $E(G)$ and defined as $HM_1(G)=\sum_{e=uv\in E(G)}d(e)^2$, where $d(e)=d(u)+d(v)$ is the edge degree. In this work we define the second hyper-Zagreb index on the adjacent edges as $HM_2(G)=\sum_{e\sim f}d(e)d(f)$, where $e\sim f$ represents the adjacent edges of $G$. By inequalities, we explore some upper and lower bounds of these hyper-Zagreb indices, and provide the relation between Zagreb indices and hyper Zagreb indices.

math.CO↗

arXiv2016

Cavitation of Water by Volume-Controlled Stretching

A liquid subjected to negative pressure is thermodynamically metastable. Confined within a small volume, negative pressure can build up until cavities form spontaneously. The critical negative pressure for cavitation in water has been theoretically predicted to be in the range of -100 to -200 MPa at room temperature, whereas values around -30 MPa have been obtained by many experiments. The discrepancy has yet to be resolved. In this study we perform molecular dynamics simulations to study cavitation of water under volume controlled stretching. It is found that liquid water exhibits a nonlinear elastic compressibility (or stretchability) under hydrostatic tension and remains stable within the confined volume until spontaneous cavitation occurs at a critical strain. Subsequently, as the volume-controlled stretching continues, the cavity grows stably and the hydrostatic tension decreases continuously until the box volume is large enough for another transition to form a water droplet. A modified nucleation theory is proposed to predict the critical condition for cavitation. In particular, a strong dependence of the critical strain and stress for cavitation on the initial liquid volume is predicted by the modified nucleation theory, which may offer a possible explanation for the discrepancies in the values of the critical negative pressure obtained from experiments.

cond-mat.soft↗

arXiv2016

On the Irregularity of Some Molecular Structures

Measures of the irregularity of chemical graphs could be helpful for QSAR/QSPR studies and for the descriptive purposes of biological and chemical properties, such as melting and boiling points, toxicity and resistance. Here we consider the following four established irregularity measures: the irregularity index by Albertson, the total irregularity, the variance of vertex degrees and the Collatz-Sinogowitz index. Through the means of graph structural analysis and derivation, we study the above-mentioned irregularity measures of several chemical molecular graphs which frequently appear in chemical, medical and material engineering, as well as the nanotubes: $TUC_4 C_8(S)$, $TUC_4 C_8(R)$, Zig-Zag $TUHC_{6}$, $TUC_4$, Armchair $TUVC_{6}$, then dendrimers $T_{k,d}$ and the circumcoronene series of benzenoid $H_k$. In addition, the irregularities of Mycielski's constructions of cycle and path graphs are analyzed.

cs.DM↗

arXiv2016

Modern Physiognomy: An Investigation on Predicting Personality Traits and Intelligence from the Human Face

The human behavior of evaluating other individuals with respect to their personality traits and intelligence by evaluating their faces plays a crucial role in human relations. These trait judgments might influence important social outcomes in our lives such as elections and court sentences. Previous studies have reported that human can make valid inferences for at least four personality traits. In addition, some studies have demonstrated that facial trait evaluation can be learned using machine learning methods accurately. In this work, we experimentally explore whether self-reported personality traits and intelligence can be predicted reliably from a facial image. More specifically, the prediction problem is separately cast in two parts: a classification task and a regression task. A facial structural feature is constructed from the relations among facial salient points, and an appearance feature is built by five texture descriptors. In addition, a minutia-based fingerprint feature from a fingerprint image is also explored. The classification results show that the personality traits "Rule-consciousness" and "Vigilance" can be predicted reliably, and that the traits of females can be predicted more accurately than those of male. However, the regression experiments show that it is difficult to predict scores for individual personality traits and intelligence. The residual plots and the correlation results indicate no evident linear correlation between the measured scores and the predicted scores. Both the classification and the regression results reveal that "Rule-consciousness" and "Tension" can be reliably predicted from the facial features, while "Social boldness" gets the worst prediction results. The experiments results show that it is difficult to predict intelligence from either the facial features or the fingerprint feature, a finding that is in agreement with previous studies.

cs.CV↗