Searcharxiv⌕ Search

arXiv subjects

Wen Li

Publications and source records attributed to Wen Li.

At least 163 records · Page 9Linked to original sources
arXiv2020

Zero-Shot Heterogeneous Transfer Learning from Recommender Systems to Cold-Start Search Retrieval

Many recent advances in neural information retrieval models, which predict top-K items given a query, learn directly from a large training set of (query, item) pairs. However, they are often insufficient when there are many previously unseen (query, item) combinations, often referred to as the cold start problem. Furthermore, the search system can be biased towards items that are frequently shown to a query previously, also known as the 'rich get richer' (a.k.a. feedback loop) problem. In light of these problems, we observed that most online content platforms have both a search and a recommender system that, while having heterogeneous input spaces, can be connected through their common output item space and a shared semantic representation. In this paper, we propose a new Zero-Shot Heterogeneous Transfer Learning framework that transfers learned knowledge from the recommender system component to improve the search component of a content platform. First, it learns representations of items and their natural-language features by predicting (item, item) correlation graphs derived from the recommender system as an auxiliary task. Then, the learned representations are transferred to solve the target search retrieval task, performing query-to-item prediction without having seen any (query, item) pairs in training. We conduct online and offline experiments on one of the world's largest search and recommender systems from Google, and present the results and lessons learned. We demonstrate that the proposed approach can achieve high performance on offline search retrieval tasks, and more importantly, achieved significant improvements on relevance and user interactions over the highly-optimized production system in online experiments.

cs.LG↗
arXiv2020

Off-Policy Reinforcement Learning for Efficient and Effective GAN Architecture Search

In this paper, we introduce a new reinforcement learning (RL) based neural architecture search (NAS) methodology for effective and efficient generative adversarial network (GAN) architecture search. The key idea is to formulate the GAN architecture search problem as a Markov decision process (MDP) for smoother architecture sampling, which enables a more effective RL-based search algorithm by targeting the potential global optimal architecture. To improve efficiency, we exploit an off-policy GAN architecture search algorithm that makes efficient use of the samples generated by previous policies. Evaluation on two standard benchmark datasets (i.e., CIFAR-10 and STL-10) demonstrates that the proposed method is able to discover highly competitive architectures for generally better image generation results with a considerably reduced computational burden: 7 GPU hours. Our code is available at https://github.com/Yuantian013/E2GAN.

cs.CV↗
arXiv2020

Analogical Image Translation for Fog Generation

Image-to-image translation is to map images from a given \emph{style} to another given \emph{style}. While exceptionally successful, current methods assume the availability of training images in both source and target domains, which does not always hold in practice. Inspired by humans' reasoning capability of analogy, we propose analogical image translation (AIT). Given images of two styles in the source domain: $\mathcal{A}$ and $\mathcal{A}^\prime$, along with images $\mathcal{B}$ of the first style in the target domain, learn a model to translate $\mathcal{B}$ to $\mathcal{B}^\prime$ in the target domain, such that $\mathcal{A}:\mathcal{A}^\prime ::\mathcal{B}:\mathcal{B}^\prime$. AIT is especially useful for translation scenarios in which training data of one style is hard to obtain but training data of the same two styles in another domain is available. For instance, in the case from normal conditions to extreme, rare conditions, obtaining real training images for the latter case is challenging but obtaining synthetic data for both cases is relatively easy. In this work, we are interested in adding adverse weather effects, more specifically fog effects, to images taken in clear weather. To circumvent the challenge of collecting real foggy images, AIT learns with synthetic clear-weather images, synthetic foggy images and real clear-weather images to add fog effects onto real clear-weather images without seeing any real foggy images during training. AIT achieves this zero-shot image translation capability by coupling a supervised training scheme in the synthetic domain, a cycle consistency strategy in the real domain, an adversarial training scheme between the two domains, and a novel network design. Experiments show the effectiveness of our method for zero-short image translation and its benefit for downstream tasks such as semantic foggy scene understanding.

cs.CV↗
arXiv2020

Deeply Aligned Adaptation for Cross-domain Object Detection

Cross-domain object detection has recently attracted more and more attention for real-world applications, since it helps build robust detectors adapting well to new environments. In this work, we propose an end-to-end solution based on Faster R-CNN, where ground-truth annotations are available for source images (e.g., cartoon) but not for target ones (e.g., watercolor) during training. Motivated by the observation that the transferabilities of different neural network layers differ from each other, we propose to apply a number of domain alignment strategies to different layers of Faster R-CNN, where the alignment strength is gradually reduced from low to higher layers. Moreover, after obtaining region proposals in our network, we develop a foreground-background aware alignment module to further reduce the domain mismatch by separately aligning features of the foreground and background regions from the source and target domains. Extensive experiments on benchmark datasets demonstrate the effectiveness of our proposed approach.

cs.CV↗
arXiv2020

A Vision of C-V2X: Technologies, Field Testing and Challenges with Chinese Development

C-V2X (Cellular Vehicle-to-Everything) is the important enabling technology for autonomous driving and intelligent transportation systems. It evolves from LTE (Long Term Evolution)-V2X to NR (New Radio)-V2X, which will coexist and be complementary with each other to provide low latency, high reliability, and high throughput communications for various C-V2X applications. In this article, a vision of C-V2X is presented. The requirements of the basic road safety and advanced applications, the architecture, the key technologies, and the standards of C-V2X are introduced, highlighting the technical evolution path from LTE-V2X to NR-V2X. Especially, based on the continual and active promotion of C-V2X research, field testing and development in China, the related works and progresses are also presented. Lastly, the trends of C-V2X applications with technical challenges are envisioned.

eess.SP↗
arXiv2020

Constructions of regular sparse anti-magic squares

Graph labeling is a well-known and intensively investigated problem in graph theory. Sparse anti-magic squares are useful in constructing vertex-magic labeling for graphs. For positive integers $n,d$ and $d<n$, an $n\times n$ array $A$ based on $\{0,1,\cdots,nd\}$ is called \emph{a sparse anti-magic square of order $n$ with density $d$}, denoted by SAMS$(n,d)$, if each element of $\{1,2,\cdots,nd\}$ occurs exactly one entry of $A$, and its row-sums, column-sums and two main diagonal sums constitute a set of $2n+2$ consecutive integers. An SAMS$(n,d)$ is called \emph{regular} if there are exactly $d$ positive entries in each row, each column and each main diagonal. In this paper, we investigate the existence of regular sparse anti-magic squares of order $n\equiv1,5\pmod 6$, and it is proved that for any $n\equiv1,5\pmod 6$, there exists a regular SAMS$(n,d)$ if and only if $2\leq d\leq n-1$.

math.CO↗
arXiv2020

OralCam: Enabling Self-Examination and Awareness of Oral Health Using a Smartphone Camera

Due to a lack of medical resources or oral health awareness, oral diseases are often left unexamined and untreated, affecting a large population worldwide. With the advent of low-cost, sensor-equipped smartphones, mobile apps offer a promising possibility for promoting oral health. However, to the best of our knowledge, no mobile health (mHealth) solutions can directly support a user to self-examine their oral health condition. This paper presents OralCam, the first interactive app that enables end-users' self-examination of five common oral conditions (diseases or early disease signals) by taking smartphone photos of one's oral cavity. OralCam allows a user to annotate additional information (e.g. living habits, pain, and bleeding) to augment the input image, and presents the output hierarchically, probabilistically and with visual explanations to help a laymen user understand examination results. Developed on our in-house dataset that consists of 3,182 oral photos annotated by dental experts, our deep learning based framework achieved an average detection sensitivity of 0.787 over five conditions with high localization accuracy. In a week-long in-the-wild user study (N=18), most participants had no trouble using OralCam and interpreting the examination results. Two expert interviews further validate the feasibility of OralCam for promoting users' awareness of oral health.

cs.HC↗
arXiv2019

Strong Performance Enhancement in Lead-Halide Perovskite Solar Cells through Rapid, Atmospheric Deposition of n-type Buffer Layer Oxides

Thin (approximately 10 nm) oxide buffer layers grown over lead-halide perovskite device stacks are critical for protecting the perovskite against mechanical and environmental damage. However, the limited perovskite stability restricts the processing methods and temperatures (<=110 C) that can be used to deposit the oxide overlayers, with the latter limiting the electronic properties of the oxides achievable. In this work, we demonstrate an alternative to existing methods that can grow pinhole-free TiOx (x = 2.00+/-0.05) films with the requisite thickness in <1 min without vacuum. This technique is atmospheric pressure chemical vapor deposition (AP-CVD). The rapid but soft deposition enables growth temperatures of >=180 °C to be used to coat the perovskite. This is >=70 °C higher than achievable by current methods and results in more conductive TiOx films, boosting solar cell efficiencies by >2%. Likewise, when AP-CVD SnOx (x ~ 2) is grown on perovskites, there is also minimal damage to the perovskite beneath. The SnOx layer is pinhole-free and conformal, which reduces shunting in devices, and increases steady-state efficiencies from 16.5% (no SnOx) to 19.4% (60 nm SnOx), with fill factors reaching 84%. This work shows AP-CVD to be a versatile technique for growing oxides on thermally-sensitive materials.

physics.app-ph↗
arXiv2019

Hybrid Organic-Metal Oxide Multilayer Channel Transistors with Record Operational Stability

Metal oxide thin-film transistors are fast becoming a ubiquitous technology for application in driving backplanes of organic light-emitting diode displays. Currently all commercial products rely on metal oxides processed via physical vapor deposition methods. Transition to simpler, higher throughput manufacturing methods such as solution-based processes, are currently been explored as cost-effective alternatives. However, developing printable oxide transistors with high carrier mobility and bias-stable operation has proved challenging. Here we show that hybrid multilayer channels composed of alternating ultra-thin layers ($\leq$4 nm) of indium oxide, zinc oxide nanoparticles, ozone-treated polystyrene and a compact zinc oxide layer, all solution-processed in ambient atmosphere, can be used to create TFTs with remarkably high electron mobility (50 cm$^{2}$/Vs) and record operational stability. Insertion of the ozone-treated polystyrene interlayer is shown to reduce the concentration of electron traps at the metal oxide surfaces and heterointerfaces. The resulting transistors exhibit dramatically enhanced bias stability over 24 h continuous operation and while subjected to large electric field flux density (2.1$\times$10$^{-6}$ C/cm$^{2}$) with no adverse effects on the electron mobility. Density functional theory calculations identify the origin of this enhanced stability as the passivation of the oxygen vacancy-related gap states due to interaction between ozonolyzed styrene moieties and the oxides. Our results sets new design guidelines for bias-stress resilient metal oxide transistors.

physics.app-ph↗
arXiv2019

Assessment of Generative Adversarial Networks Model for Synthetic Optical Coherence Tomography Images of Retinal Disorders

Purpose: To assess whether a generative adversarial network (GAN) could synthesize realistic optical coherence tomography (OCT) images that satisfactorily serve as the educational images for retinal specialists and the training datasets for the classification of various retinal disorders using deep learning (DL). Methods: The GANs architecture was adopted to synthesis high-resolution OCT images training on a publicly available OCT dataset including urgent referrals (choroidal neovascularization and diabetic macular edema) and non-urgent referrals (normal and drusen). 400 real and synthetic OCT images were evaluated by 2 retinal specialists to assess image quality. We further trained 2 DL models on either real or synthetic datasets and compared the performance of urgent vs nonurgent referrals diagnosis tested on a local (1000 images from the public dataset) and clinical validation dataset (278 images from Shanghai Shibei Hospital). Results: The image quality of real vs synthetic OCT images was similar as assessed by 2 retinal specialists. The accuracy of discrimination as real vs synthetic OCT images was 59.50% for retinal specialist 1 and 53.67% for retinal specialist 2. For the local dataset, the DL model trained on real (DL_Model_R) and synthetic OCT images (DL_Model_S) had an area under the curve (AUC) of 0.99, and 0.98 respectively. For the clinical dataset, the AUC was 0.94 for DL_Model_R, 0.90 for DL_Model_S. Conclusions: The GAN-synthetic OCT images can be used by clinicians for educational purposes and developing DL algorithms. Translational Relevance: The medical image synthesis based on GANs is promising in human and machine to fulfill clinical tasks.

physics.med-ph↗
arXiv2019

Resonant excitation of whistler waves by a helical electron beam

Chorus-like whistler-mode waves that are known to play a fundamental role in driving radiation-belt dynamics are excited on the Large Plasma Device by the injection of a helical electron beam into a cold plasma. The mode structure of the excited whistler wave is identified using a phase-correlation technique showing that the waves are excited through a combination of Landau resonance, cyclotron resonance and anomalous cyclotron resonance. The dominant wave mode excited through cyclotron resonance is quasi-parallel propagating, whereas wave modes excited through Landau resonance and anomalous cyclotron resonance propagate at oblique angles that are close to the resonance cone. An analysis of the linear wave growth rates captures the major observations in the experiment. The results have important implications for the generation process of whistler waves in the Earth's inner magnetosphere.

physics.plasm-ph↗
arXiv2019

Semi-Supervised Learning by Augmented Distribution Alignment

In this work, we propose a simple yet effective semi-supervised learning approach called Augmented Distribution Alignment. We reveal that an essential sampling bias exists in semi-supervised learning due to the limited number of labeled samples, which often leads to a considerable empirical distribution mismatch between labeled data and unlabeled data. To this end, we propose to align the empirical distributions of labeled and unlabeled data to alleviate the bias. On one hand, we adopt an adversarial training strategy to minimize the distribution distance between labeled and unlabeled data as inspired by domain adaptation works. On the other hand, to deal with the small sample size issue of labeled data, we also propose a simple interpolation strategy to generate pseudo training samples. Those two strategies can be easily implemented into existing deep neural networks. We demonstrate the effectiveness of our proposed approach on the benchmark SVHN and CIFAR10 datasets. Our code is available at \url{https://github.com/qinenergy/adanet}.

cs.CV↗
arXiv2019

DLOW: Domain Flow for Adaptation and Generalization

In this work, we present a domain flow generation(DLOW) model to bridge two different domains by generating a continuous sequence of intermediate domains flowing from one domain to the other. The benefits of our DLOW model are two-fold. First, it is able to transfer source images into different styles in the intermediate domains. The transferred images smoothly bridge the gap between source and target domains, thus easing the domain adaptation task. Second, when multiple target domains are provided for training, our DLOW model is also able to generate new styles of images that are unseen in the training data. We implement our DLOW model based on CycleGAN. A domainness variable is introduced to guide the model to generate the desired intermediate domain images. In the inference phase, a flow of various styles of images can be obtained by varying the domainness variable. We demonstrate the effectiveness of our model for both cross-domain semantic segmentation and the style generalization tasks on benchmark datasets. Our implementation is available at https://github.com/ETHRuiGong/DLOW.

cs.CV↗
arXiv2019

Known-class Aware Self-ensemble for Open Set Domain Adaptation

Existing domain adaptation methods generally assume different domains have the identical label space, which is quite restrict for real-world applications. In this paper, we focus on a more realistic and challenging case of open set domain adaptation. Particularly, in open set domain adaptation, we allow the classes from the source and target domains to be partially overlapped. In this case, the assumption of conventional distribution alignment does not hold anymore, due to the different label spaces in two domains. To tackle this challenge, we propose a new approach coined as Known-class Aware Self-Ensemble (KASE), which is built upon the recently developed self-ensemble model. In KASE, we first introduce a Known-class Aware Recognition (KAR) module to identify the known and unknown classes from the target domain, which is achieved by encouraging a low cross-entropy for known classes and a high entropy based on the source data from the unknown class. Then, we develop a Known-class Aware Adaptation (KAA) module to better adapt from the source domain to the target by reweighing the adaptation loss based on the likeliness to belong to known classes of unlabeled target samples as predicted by KAR. Extensive experiments on multiple benchmark datasets demonstrate the effectiveness of our approach.

cs.CV↗
arXiv2019

Sliced Wasserstein Generative Models

In generative modeling, the Wasserstein distance (WD) has emerged as a useful metric to measure the discrepancy between generated and real data distributions. Unfortunately, it is challenging to approximate the WD of high-dimensional distributions. In contrast, the sliced Wasserstein distance (SWD) factorizes high-dimensional distributions into their multiple one-dimensional marginal distributions and is thus easier to approximate. In this paper, we introduce novel approximations of the primal and dual SWD. Instead of using a large number of random projections, as it is done by conventional SWD approximation methods, we propose to approximate SWDs with a small number of parameterized orthogonal projections in an end-to-end deep learning fashion. As concrete applications of our SWD approximations, we design two types of differentiable SWD blocks to equip modern generative frameworks---Auto-Encoders (AE) and Generative Adversarial Networks (GAN). In the experiments, we not only show the superiority of the proposed generative models on standard image synthesis benchmarks, but also demonstrate the state-of-the-art performance on challenging high resolution image and video generation in an unsupervised manner.

cs.CV↗
arXiv2019

Sliced Wasserstein Generative Models

In generative modeling, the Wasserstein distance (WD) has emerged as a useful metric to measure the discrepancy between generated and real data distributions. Unfortunately, it is challenging to approximate the WD of high-dimensional distributions. In contrast, the sliced Wasserstein distance (SWD) factorizes high-dimensional distributions into their multiple one-dimensional marginal distributions and is thus easier to approximate. In this paper, we introduce novel approximations of the primal and dual SWD. Instead of using a large number of random projections, as it is done by conventional SWD approximation methods, we propose to approximate SWDs with a small number of parameterized orthogonal projections in an end-to-end deep learning fashion. As concrete applications of our SWD approximations, we design two types of differentiable SWD blocks to equip modern generative frameworks---Auto-Encoders (AE) and Generative Adversarial Networks (GAN). In the experiments, we not only show the superiority of the proposed generative models on standard image synthesis benchmarks, but also demonstrate the state-of-the-art performance on challenging high resolution image and video generation in an unsupervised manner.

cs.CV↗
arXiv2019

Finite element based Green's function integral equation for modelling light scattering

We revisit the volume Green's function integral equation for modelling light scattering with discretization strategies as well as numerical integration recipes borrowed from finite element method. The merits of introducing finite element techniques into Green's function integral equation are apparent. Firstly, the finite element discretization provides a much better geometric approximation of the scatters, compared with that of the conventional discretization method using staircase approximation. Secondly, the accuracy of numerical integral inside one element associated with all different types of Green's function integral equations can be greatly improved by introducing quadrature rules. Within the standard framework of Green's function integral equation, we seamlessly introduce finite element techniques into the Green's function integral equation by introducing the auxiliary variables that confines the singular integrand inside each element, leading to a better and more flexible approximation of the geometry of the scatters and a more accurate numerical integral. We then illustrate the advantages of our finite element based Green's function integral equation method via a few concrete examples in modelling light scattering by optically large and complex scatters.

physics.optics↗
arXiv2019

On the statistics of acceleration and loss of relativistic electrons in the outer radiation belt: a superposed epoch analysis

We investigate the response of the outer Van Allen belt electrons to various types of solar wind and magnetospheric disturbances. We use electron phase space density (PSD) calculations as well as concurrent Pc5 and chorus wave activity observations in the outer belt during the Van Allen Probes era to compare 20 electron enhancement and 8 depletion events. Results indicate that the combined effect of magnetopause shadowing and outward diffusion driven by Pc5 waves is present in both groups of events. Furthermore, in the case of enhancement events, the synergy of enhanced seed population levels and chorus activity -- due to the enhanced substorm activity -- can effectively replenish the losses of relativistic electrons, while inward diffusion can further accelerate them.

physics.space-ph↗