arXiv subjects

Wenbin Li

Publications and source records attributed to Wenbin Li.

At least 145 records · Page 8Linked to original sources

arXiv2020

Diversity Helps: Unsupervised Few-shot Learning via Distribution Shift-based Data Augmentation

Few-shot learning aims to learn a new concept when only a few training examples are available, which has been extensively explored in recent years. However, most of the current works heavily rely on a large-scale labeled auxiliary set to train their models in an episodic-training paradigm. Such a kind of supervised setting basically limits the widespread use of few-shot learning algorithms. Instead, in this paper, we develop a novel framework called Unsupervised Few-shot Learning via Distribution Shift-based Data Augmentation (ULDA), which pays attention to the distribution diversity inside each constructed pretext few-shot task when using data augmentation. Importantly, we highlight the value and importance of the distribution diversity in the augmentation-based pretext few-shot tasks, which can effectively alleviate the overfitting problem and make the few-shot model learn more robust feature representations. In ULDA, we systemically investigate the effects of different augmentation techniques and propose to strengthen the distribution diversity (or difference) between the query set and support set in each few-shot task, by augmenting these two sets diversely (i.e., distribution shifting). In this way, even incorporated with simple augmentation techniques (e.g., random crop, color jittering, or rotation), our ULDA can produce a significant improvement. In the experiments, few-shot models learned by ULDA can achieve superior generalization performance and obtain state-of-the-art results in a variety of established few-shot learning tasks on Omniglot and miniImageNet. The source code is available in https://github.com/WonderSeven/ULDA.

cs.CV↗

arXiv2020

Embedded Deep Bilinear Interactive Information and Selective Fusion for Multi-view Learning

As a concrete application of multi-view learning, multi-view classification improves the traditional classification methods significantly by integrating various views optimally. Although most of the previous efforts have been demonstrated the superiority of multi-view learning, it can be further improved by comprehensively embedding more powerful cross-view interactive information and a more reliable multi-view fusion strategy in intensive studies. To fulfill this goal, we propose a novel multi-view learning framework to make the multi-view classification better aimed at the above-mentioned two aspects. That is, we seamlessly embed various intra-view information, cross-view multi-dimension bilinear interactive information, and a new view ensemble mechanism into a unified framework to make a decision via the optimization. In particular, we train different deep neural networks to learn various intra-view representations, and then dynamically learn multi-dimension bilinear interactive information from different bilinear similarities via the bilinear function between views. After that, we adaptively fuse the representations of multiple views by flexibly tuning the parameters of the view-weight, which not only avoids the trivial solution of weight but also provides a new way to select a few discriminative views that are beneficial to make a decision for the multi-view classification. Extensive experiments on six publicly available datasets demonstrate the effectiveness of the proposed method.

cs.CV↗

arXiv2020

Experimental evidence of monolayer AlB$_2$ with symmetry-protected Dirac cones

Monolayer AlB$_2$ is composed of two atomic layers: honeycomb borophene and triangular aluminum. In contrast with the bulk phase, monolayer AlB$_2$ is predicted to be a superconductor with a high critical temperature. Here, we demonstrate that monolayer AlB$_2$ can be synthesized on Al(111) via molecular beam epitaxy. Our theoretical calculations revealed that the monolayer AlB$_2$ hosts several Dirac cones along the $Γ$--M and $Γ$--K directions; these Dirac cones are protected by crystal symmetries and are thus resistant to external perturbations. The extraordinary electronic structure of the monolayer AlB$_2$ was confirmed via angle-resolved photoemission spectroscopy measurements. These results are likely to stimulate further research interest to explore the exotic properties arising from the interplay of Dirac fermions and superconductivity in two-dimensional materials.

cond-mat.mtrl-sci↗

arXiv2020

Alleviating the Incompatibility between Cross Entropy Loss and Episode Training for Few-shot Skin Disease Classification

Skin disease classification from images is crucial to dermatological diagnosis. However, identifying skin lesions involves a variety of aspects in terms of size, color, shape, and texture. To make matters worse, many categories only contain very few samples, posing great challenges to conventional machine learning algorithms and even human experts. Inspired by the recent success of Few-Shot Learning (FSL) in natural image classification, we propose to apply FSL to skin disease identification to address the extreme scarcity of training sample problem. However, directly applying FSL to this task does not work well in practice, and we find that the problem can be largely attributed to the incompatibility between Cross Entropy (CE) and episode training, which are both commonly used in FSL. Based on a detailed analysis, we propose the Query-Relative (QR) loss, which proves superior to CE under episode training and is closely related to recently proposed mutual information estimation. Moreover, we further strengthen the proposed QR loss with a novel adaptive hard margin strategy. Comprehensive experiments validate the effectiveness of the proposed FSL scheme and the possibility to diagnosis rare skin disease with a few labeled samples.

eess.IV↗

arXiv2020

RGBD-Dog: Predicting Canine Pose from RGBD Sensors

The automatic extraction of animal \reb{3D} pose from images without markers is of interest in a range of scientific fields. Most work to date predicts animal pose from RGB images, based on 2D labelling of joint positions. However, due to the difficult nature of obtaining training data, no ground truth dataset of 3D animal motion is available to quantitatively evaluate these approaches. In addition, a lack of 3D animal pose data also makes it difficult to train 3D pose-prediction methods in a similar manner to the popular field of body-pose prediction. In our work, we focus on the problem of 3D canine pose estimation from RGBD images, recording a diverse range of dog breeds with several Microsoft Kinect v2s, simultaneously obtaining the 3D ground truth skeleton via a motion capture system. We generate a dataset of synthetic RGBD images from this data. A stacked hourglass network is trained to predict 3D joint locations, which is then constrained using prior models of shape and pose. We evaluate our model on both synthetic and real RGBD images and compare our results to previously published work fitting canine models to images. Finally, despite our training set consisting only of dog data, visual inspection implies that our network can produce good predictions for images of other quadrupeds -- e.g. horses or cats -- when their pose is similar to that contained in our training set.

cs.CV↗

arXiv2020

Semantic Regularization: Improve Few-shot Image Classification by Reducing Meta Shift

Few-shot image classification requires the classifier to robustly cope with unseen classes even if there are only a few samples for each class. Recent advances benefit from the meta-learning process where episodic tasks are formed to train a model that can adapt to class change. However, these task sare independent to each other and existing works mainly rely on limited samples of individual support set in a single meta task. This strategy leads to severe meta shift issues across multiple tasks, meaning the learned prototypes or class descriptors are not stable as each task only involves their own support set. To avoid this problem, we propose a concise Semantic RegularizationNetwork to learn a common semantic space under the framework of meta-learning. In this space, all class descriptors can be regularized by the learned semantic basis, which can effectively solve the meta shift problem. The key is to train a class encoder and decoder structure that can encode the sample embedding features into the semantic domain with trained semantic basis, and generate a more stable and general class descriptor from the decoder. We evaluate our work by extensive comparisons with previous methods on three benchmark datasets (MiniImageNet, TieredImageNet, and CUB). The results show that the semantic regularization module improves performance by 4%-7% over the baseline method, and achieves competitive results over the current state-of-the-art models.

cs.CV↗

arXiv2020

Asymmetric Distribution Measure for Few-shot Learning

The core idea of metric-based few-shot image classification is to directly measure the relations between query images and support classes to learn transferable feature embeddings. Previous work mainly focuses on image-level feature representations, which actually cannot effectively estimate a class's distribution due to the scarcity of samples. Some recent work shows that local descriptor based representations can achieve richer representations than image-level based representations. However, such works are still based on a less effective instance-level metric, especially a symmetric metric, to measure the relations between query images and support classes. Given the natural asymmetric relation between a query image and a support class, we argue that an asymmetric measure is more suitable for metric-based few-shot learning. To that end, we propose a novel Asymmetric Distribution Measure (ADM) network for few-shot learning by calculating a joint local and global asymmetric measure between two multivariate local distributions of queries and classes. Moreover, a task-aware Contrastive Measure Strategy (CMS) is proposed to further enhance the measure function. On popular miniImageNet and tieredImageNet, we achieve $3.02\%$ and $1.56\%$ gains over the state-of-the-art method on the $5$-way $1$-shot task, respectively, validating our innovative design of asymmetric distribution measures for few-shot learning.

cs.CV↗

arXiv2019

First-principles calculations of charge carrier mobility and conductivity in bulk semiconductors and two-dimensional materials

One of the fundamental properties of semiconductors is their ability to support highly tunable electric currents in the presence of electric fields or carrier concentration gradients. These properties are described by transport coefficients such as electron and hole mobilities. Recently, advances in electronic structure methods for real materials have made it possible to study these properties with predictive accuracy and without resorting to empirical parameters. Here, we review the most recent developments in the area of ab initio calculations of carrier mobilities of semiconductors. In the first part, we offer a brief historical overview of approaches to the calculation of carrier mobilities, and we establish the conceptual framework underlying modern ab initio approaches. We summarize the Boltzmann theory of carrier transport and we discuss its scope of applicability, merits, and limitations in the broader context of many-body Green's function approaches. We discuss recent implementations of the Boltzmann formalism within the context of density functional theory and many-body perturbation theory calculations, placing an emphasis on the key computational challenges and suggested solutions. In the second part, we discuss recent investigations of classic materials such as silicon, diamond, GaAs, GaN, Ga2O3, and lead halide perovskites as well as low-dimensional semiconductors such as graphene, silicene, phosphorene, MoS2, and InSe. We also review recent efforts toward high-throughput calculations of carrier transport. In the last part, we discuss the extension of the methodology to study spintronics and topological materials and we comment on the possibility of incorporating Berry-phase effects and many-body correlations beyond the standard Boltzmann formalism.

cond-mat.mtrl-sci↗

arXiv2019

Learning Manipulation under Physics Constraints with Visual Perception

Understanding physical phenomena is a key competence that enables humans and animals to act and interact under uncertain perception in previously unseen environments containing novel objects and their configurations. In this work, we consider the problem of autonomous block stacking and explore solutions to learning manipulation under physics constraints with visual perception inherent to the task. Inspired by the intuitive physics in humans, we first present an end-to-end learning-based approach to predict stability directly from appearance, contrasting a more traditional model-based approach with explicit 3D representations and physical simulation. We study the model's behavior together with an accompanied human subject test. It is then integrated into a real-world robotic system to guide the placement of a single wood block into the scene without collapsing existing tower structure. To further automate the process of consecutive blocks stacking, we present an alternative approach where the model learns the physics constraint through the interaction with the environment, bypassing the dedicated physics learning as in the former part of this work. In particular, we are interested in the type of tasks that require the agent to reach a given goal state that may be different for every new trial. Thereby we propose a deep reinforcement learning framework that learns policies for stacking tasks which are parametrized by a target structure.

cs.RO↗

arXiv2019

Revisiting Local Descriptor based Image-to-Class Measure for Few-shot Learning

Few-shot learning in image classification aims to learn a classifier to classify images when only few training examples are available for each class. Recent work has achieved promising classification performance, where an image-level feature based measure is usually used. In this paper, we argue that a measure at such a level may not be effective enough in light of the scarcity of examples in few-shot learning. Instead, we think a local descriptor based image-to-class measure should be taken, inspired by its surprising success in the heydays of local invariant features. Specifically, building upon the recent episodic training mechanism, we propose a Deep Nearest Neighbor Neural Network (DN4 in short) and train it in an end-to-end manner. Its key difference from the literature is the replacement of the image-level feature based measure in the final layer by a local descriptor based image-to-class measure. This measure is conducted online via a $k$-nearest neighbor search over the deep local descriptors of convolutional feature maps. The proposed DN4 not only learns the optimal deep local descriptors for the image-to-class measure, but also utilizes the higher efficiency of such a measure in the case of example scarcity, thanks to the exchangeability of visual patterns across the images in the same class. Our work leads to a simple, effective, and computationally efficient framework for few-shot learning. Experimental study on benchmark datasets consistently shows its superiority over the related state-of-the-art, with the largest absolute improvement of $17\%$ over the next best. The source code can be available from \UrlFont{https://github.com/WenbinLee/DN4.git}.

cs.CV↗

arXiv2019

MID-Fusion: Octree-based Object-Level Multi-Instance Dynamic SLAM

We propose a new multi-instance dynamic RGB-D SLAM system using an object-level octree-based volumetric representation. It can provide robust camera tracking in dynamic environments and at the same time, continuously estimate geometric, semantic, and motion properties for arbitrary objects in the scene. For each incoming frame, we perform instance segmentation to detect objects and refine mask boundaries using geometric and motion information. Meanwhile, we estimate the pose of each existing moving object using an object-oriented tracking method and robustly track the camera pose against the static scene. Based on the estimated camera pose and object poses, we associate segmented masks with existing models and incrementally fuse corresponding colour, depth, semantic, and foreground object probabilities into each object model. In contrast to existing approaches, our system is the first system to generate an object-level dynamic volumetric map from a single RGB-D camera, which can be used directly for robotic tasks. Our method can run at 2-3 Hz on a CPU, excluding the instance segmentation part. We demonstrate its effectiveness by quantitatively and qualitatively testing it on both synthetic and real-world sequences.

cs.RO↗

arXiv2018

CariGAN: Caricature Generation through Weakly Paired Adversarial Learning

Caricature generation is an interesting yet challenging task. The primary goal is to generate plausible caricatures with reasonable exaggerations given face images. Conventional caricature generation approaches mainly use low-level geometric transformations such as image warping to generate exaggerated images, which lack richness and diversity in terms of content and style. The recent progress in generative adversarial networks (GANs) makes it possible to learn an image-to-image transformation from data, so that richer contents and styles can be generated. However, directly applying the GAN-based models to this task leads to unsatisfactory results because there is a large variance in the caricature distribution. Moreover, some models require strictly paired training data which largely limits their usage scenarios. In this paper, we propose CariGAN overcome these problems. Instead of training on paired data, CariGAN learns transformations only from weakly paired images. Specifically, to enforce reasonable exaggeration and facial deformation, facial landmarks are adopted as an additional condition to constrain the generated image. Furthermore, an attention mechanism is introduced to encourage our model to focus on the key facial parts so that more vivid details in these regions can be generated. Finally, a Diversity Loss is proposed to encourage the model to produce diverse results to help alleviate the `mode collapse' problem of the conventional GAN-based models. Extensive experiments on a new large-scale `WebCaricature' dataset show that the proposed CariGAN can generate more plausible caricatures with larger diversity compared with the state-of-the-art models.

cs.CV↗

arXiv2018

InteriorNet: Mega-scale Multi-sensor Photo-realistic Indoor Scenes Dataset

Datasets have gained an enormous amount of popularity in the computer vision community, from training and evaluation of Deep Learning-based methods to benchmarking Simultaneous Localization and Mapping (SLAM). Without a doubt, synthetic imagery bears a vast potential due to scalability in terms of amounts of data obtainable without tedious manual ground truth annotations or measurements. Here, we present a dataset with the aim of providing a higher degree of photo-realism, larger scale, more variability as well as serving a wider range of purposes compared to existing datasets. Our dataset leverages the availability of millions of professional interior designs and millions of production-level furniture and object assets -- all coming with fine geometric details and high-resolution texture. We render high-resolution and high frame-rate video sequences following realistic trajectories while supporting various camera types as well as providing inertial measurements. Together with the release of the dataset, we will make executable program of our interactive simulator software as well as our renderer available at https://interiornetdataset.github.io. To showcase the usability and uniqueness of our dataset, we show benchmarking results of both sparse and dense SLAM algorithms.

cs.CV↗

arXiv2018

WebCaricature: a benchmark for caricature recognition

Studying caricature recognition is fundamentally important to understanding of face perception. However, little research has been conducted in the computer vision community, largely due to the shortage of suitable datasets. In this paper, a new caricature dataset is built, with the objective to facilitate research in caricature recognition. All the caricatures and face images were collected from the Web. Compared with two existing datasets, this dataset is much more challenging, with a much greater number of available images, artistic styles and larger intra-personal variations. Evaluation protocols are also offered together with their baseline performances on the dataset to allow fair comparisons. Besides, a framework for caricature face recognition is presented to make a thorough analyze of the challenges of caricature recognition. By analyzing the challenges, the goal is to show problems that worth to be further investigated. Additionally, based on the evaluation protocols and the framework, baseline performances of various state-of-the-art algorithms are provided. A conclusion is that there is still a large space for performance improvement and the analyzed problems still need further investigation.

cs.CV↗

arXiv2017

Experimental realization of graphene-like borophene

We report a successful preparation of a purely honeycomb, graphene-like borophene, by using an Al(111) surface as the substrate and molecular beam epitaxy (MBE) growth in ultrahigh vacuum. Our scanning tunneling microscopy (STM) reveals perfect monolayer borophene with planar, non-buckled honeycomb lattice similar as graphene. Theoretical calculations show that the honeycomb borophene on Al(111) is energetically stable. Remarkably, nearly one electron charge is transferred to each boron atom from the Al(111) substrate and stabilizes the honeycomb borophene structure, in contrast to the little charge transfer in borophene/Ag(111) case. This work demonstrates that the borophene lattice can be manipulated by controlling the charge transfer between the substrate and the borophene film.

cond-mat.mes-hall↗

arXiv2017

Direct evidences of pentagonal silicon chains and magic clusters

Pentagon is one of the most beautiful geometric structures in nature, but it is rarely seen simply because five-fold symmetry is mathematically forbidden in a 2D or 3D periodic lattices. Fortunately, pentagon as a structural element is allowed in 1D or 0D systems, since translational symmetry is not necessary there. However, in these systems pentagons only compose a small portion of the structure. So far, 1D or 0D systems consisting of purely pentagons are still rare. Here, combing high resolution non-contact atomic force microscopy and tip-enhanced Raman spectroscopy, we have directly visualized the pentagon-ring structure in self-assembled Si nanoribbons and magic clusters on Ag(110) substrate. Moreover, chemical fingerprint of Si pentagon was detected in individual Si nanoribbon and clusters by tip-enhanced Raman spectroscopy. This work demonstrates that Si pentagon can be an important element in building silicon nanostructures, which may find potential applications in future nanoelectronics based on silicon.

cond-mat.mes-hall↗

arXiv2017

Acquiring Target Stacking Skills by Goal-Parameterized Deep Reinforcement Learning

Understanding physical phenomena is a key component of human intelligence and enables physical interaction with previously unseen environments. In this paper, we study how an artificial agent can autonomously acquire this intuition through interaction with the environment. We created a synthetic block stacking environment with physics simulation in which the agent can learn a policy end-to-end through trial and error. Thereby, we bypass to explicitly model physical knowledge within the policy. We are specifically interested in tasks that require the agent to reach a given goal state that may be different for every new trial. To this end, we propose a deep reinforcement learning framework that learns policies which are parametrized by a goal. We validated the model on a toy example navigating in a grid world with different target positions and in a block stacking task with different target structures of the final tower. In contrast to prior work, our policies show better generalization across different goals.

cs.RO↗

arXiv2017

Automatic Structural Scene Digitalization

In this paper, we present an automatic system for the analysis and labeling of structural scenes, floor plan drawings in Computer-aided Design (CAD) format. The proposed system applies a fusion strategy to detect and recognize various components of CAD floor plans, such as walls, doors, windows and other ambiguous assets. Technically, a general rule-based filter parsing method is fist adopted to extract effective information from the original floor plan. Then, an image-processing based recovery method is employed to correct information extracted in the first step. Our proposed method is fully automatic and real-time. Such analysis system provides high accuracy and is also evaluated on a public website that, on average, archives more than ten thousands effective uses per day and reaches a relatively high satisfaction rate.

cs.GR↗