Designing efficient deep learning models for 3D point cloud perception is becoming a major research direction. Point-voxel convolution (PVConv) Liu et al. (2019) is a pioneering research work in this ...topic. However, since with quite a few layers of simple 3D convolutions and linear point-voxel feature fusion operations, it still has considerable room for improvement in performance. In this paper, we propose a novel pyramid point-voxel convolution (PyraPVConv) block with two key structural modifications to address the above issues. First, PyraPVConv uses a voxel pyramid module to fully extract voxel features in the manner of feature pyramid, such that sufficient voxel features can be obtained efficiently. Second, a sharable attention module is utilized to capture compatible features between multi-scale voxels in pyramid and point cloud for aggregation, as well as to reduce the complexity via structure sharing. Extensive results on three point cloud perception tasks, i.e., indoor scene segmentation, object part segmentation and 3D object detection, validate that the networks constructed by stacking PyraPVConv blocks are efficient in terms of both GPU memory consumption and computational complexity, and are superior to the state-of-the-art methods.
Adversarial attacks have been successfully extended to the field of point clouds. Besides applying the common perturbation guided by the gradient, adversarial attacks on point clouds can be conducted ...by applying directional perturbations, e.g., along normal and along the tangent plane. In this article, we first investigate whether adversarial attacks with these two orthogonal directional perturbations are more imperceptible than that with the gradient-aware perturbation. Second, we investigate the deeper difference between adversarial attacks with these two directional perturbations, and whether they are applicable to the same scenarios. Third, based on the verification results that the above two directional perturbations have different sensitiveness to curvature, we devise a novel normal-tangent attack (NTA) framework with a hybrid directional perturbation scheme that adaptively chooses the direction according to the curvature of the local shape around the point. Extensive experiments on two publicly available data sets, e.g., ModelNet40 and ShapeNet Part, with classifiers in three representative networks, e.g., PointNet++, DGCNN, PointConv, validate the effectiveness of NTA, and the superiority to the state-of-the-art methods.
Graph neural network (GNN) has achieved remarkable success in various graph learning tasks, such as node classification, link prediction, and graph classification. The key to the success of GNN lies ...in its effective structure information representation through neighboring aggregation. However, the attacker can easily perturb the aggregation process through injecting fake nodes, which reveals that GNN is vulnerable to the graph injection attack (GIA). Existing GIA methods primarily focus on damaging the classical feature aggregation process while overlooking the neighborhood aggregation process via label propagation. To bridge this gap, we propose the label-propagation-based global injection attack (LPGIA) which conducts the GIA on the node classification task. Specifically, we analyze the aggregation process from the perspective of label propagation and transform the GIA problem into a global injection label specificity attack problem. To solve this problem, LPGIA utilizes a label-propagation-based strategy to optimize the combinations of the nodes connected to the injected node. Then, LPGIA leverages the feature mapping to generate malicious features for injected nodes. In extensive experiments against representative GNNs, LPGIA outperforms the previous best-performing injection attack method in various datasets, demonstrating its superiority and transferability.
It is a challenging task to handle the vector field visualization at local critical points. Generally, topological based methods firstly divide critical regions into different categories, and then ...process the different types of critical regions to improve the effect, which pipeline is complex. In the paper, a learning based multi-task super resolution (SR) method is proposed to improve the refinement of vector field, and enhance the visualization effect, especially at the critical region. In detail, the multi-task model consists of two important des igns on task branches: one task is to simulate the interpolation of discrete vector fields based on an improved super-resolution network; and the other is a classification task to identify the types of critical vector fields. It is an efficient end-to-end architecture for both training and inferencing stages, which simplifies the pipeline of critical vector field visualization and improves the visualization effect. In experiment, we compare our method with both traditional interpolation and pure SR network on both simulation data and real data, and the reported results indicate our method lower the error and improve PSNR significantly.
It is a challenging task to handle the vector field visualization at local critical points. Generally, topological based methods firstly divide critical regions into different categories, and then ...process the different types of critical regions to improve the effect, which pipeline is complex. In the paper, a learning based multi-task super resolution (SR) method is proposed to improve the refinement of vector field, and enhance the visualization effect, especially at the critical region. In detail, the multi-task model consists of two important des igns on task branches: one task is to simulate the interpolation of discrete vector fields based on an improved super-resolution network; and the other is a classification task to identify the types of critical vector fields. It is an efficient end-to-end architecture for both training and inferencing stages, which simplifies the pipeline of critical vector field visualization and improves the visualization effect. In experiment, we compare our method with both traditional interpolation and pure SR network on both simulation data and real data, and the reported results indicate our method lower the error and improve PSNR significantly.
Many efforts have been made on developing adversarial attack methods on point clouds. However, without fully considering the geometric property of point clouds, existing methods tend to produce ...clearly visible outliers. In this paper, we propose a novel NormalAttack framework towards imperceptible adversarial attacks on point clouds. First, we enforce the perturbation to be concentrated along normals to deform the underlying surface of 3D point clouds, such that tiny perturbation can make the shape deformed for better attack performance. Second, we guide the perturbation to be located more on regions with larger curvature, such that better imperceptibility is achieved. Extensive experiments on three representative networks, e.g., PointNet++, DGCNN, and PointConv, validate the effectiveness of NormalAttack and its superiority to state-of-the-art methods.
The paper is aimed to investigate the effect of lamellar structure distribution around elliptical notch in bimodal titanium alloys of Ti-6Al-4V. In this regard, within the framework of crystal ...plasticity based finite element (CPFE), Representative volume element (RVE) model with elliptical notch for bimodal Ti–6Al–4V is developed, to address the effect of lamellar structure on fatigue resistance performance under external fatigue loading. Lamellar structure distribution around elliptical notches with different aspect ratios are investigated. Strain along radial and circular paths are extracted to identify the possible crack initiation points. The numerical results show that, Strain distribution around notches can be adjusted through lamellar structure distribution such that strain uniformity is achieved. Increase of strain oscillation caused by lamellar structure leads to the improvement of fatigue resistance. With the increase of notch aspect ratio, elliptical notch eventually evolves into line crack and tuning performance of lamellar structure appreciably declines. This numerical study provides a new insight into lamellar structure tuning in dual-phase titanium alloys.
Metamaterials, rationally designed multiscale composite systems, have attracted extensive interest because of their potential for a broad range of applications due to their unique properties such as ...negative Poisson's ratio, exceptional mechanical performance, tunable photonic and phononic properties, structural reconfiguration,
etc.
Though they are dominated by an auxetic structure, the constituents of metamaterials also play an indispensable role in determining their unprecedented properties. In this vein, 2D materials such as graphene, silicene, and phosphorene with superior structural tunability are ideal candidates for constituents of metamaterials. However, the nanostructure-property relationship and composition-property relationship of these 2D material-based metamaterials remain largely unexplored. Mechanical anisotropy inherited from the 2D material constituents, for example, may substantially impact the physical stability and robustness of the corresponding metamaterial systems. Herein, classical molecular dynamics simulations are performed using a generic coarse-grained model to explore the deformation mechanism of these 2D material-based metamaterials with sinusoidally curved ligaments and the effect of mechanical anisotropy on mechanical properties, especially the negative Poisson's ratio. The results indicate that deformation under axial tensile load can be divided into two stages: bending-dominated and stretching-dominated, in which the rotation of junctions in the former stage results in auxetic behavior of the proposed metamaterials. In addition, the auxetic behavior depends heavily on both the amplitude/wavelength ratio of the sinusoidal ligament and the stiffness ratio between axial and transverse directions. The magnitude of negative Poisson's ratio increases from 0 to 0.625, with an associated increase of the amplitude/wavelength ratio from 0 to 0.225, and fluctuates at around 0.625, in good agreement with the literature, with amplitude/wavelength ratios greater than 0.225. More interestingly, the magnitude of negative Poisson's ratio increases from 0.47 to 0.87 with the increase of the stiffness ratio from 0.125 to 8, in good agreement with additional all-atom molecular dynamics simulations for phosphorene and molybdenum disulfide. Overall, these research findings shed light on the deformation mechanism of auxetic metamaterials, providing useful guidelines for designing auxetic 2D lattice structures made of 2D materials that can display a tunable negative Poisson's ratio.
Mechanical properties, especially negative Poisson's, of 2D sinusoidal lattice metamaterials based on 2D materials depends highly on both geometrical factors and tuned mechanical anisotropy according to our generic coarse-grained molecular dynamics simulations.
Adversarial training (AT) is one of the most effective ways against adversarial attacks. However, multi-step AT is time-consuming while single-step AT is ineffective. In this paper, we propose an ...Energy-AT framework to make single-step AT as effective as multi-step ones, by exploiting the two properties of energy-based models (EBM). First, we utilize the Helmholtz free energy in EBM to push generated examples to be outside of the distribution boundaries of their categories, such that they are more adversarial. Second, we apply an adaptive temperature scheme in EBM to amplify the training gradients of weak adversarial examples targetedly, such that those originally hard-to-learn examples contribute to the robustification of models also. Extensive experiments validate that Energy-AT improves the robustness of models significantly to adversarial attacks in both white-box and black-box settings, and outperforms the state-of-the-art methods.