The crack-damage resistance of early-age concrete is affected by multiple factors such as hydration, self-drying, temperature and humidity diffusion, and material properties, which are difficult to be accurately evaluated by traditional theories and numerical models. This paper proposed an adaptive physics-informed back propagation neural network (BPINN) to accurately evaluate the damage of early-age concrete under multi-physics field coupling. The temperature and humidity diffusion and shrinkage models are used as physics loss functions to guide the model in learning the physics laws. Furthermore, time-dependent factor weights are constructed for both the physics and boundary equations to enhance the model's ability to learn the spatiotemporal feature distribution of the sampling points. BPINN effectively simulates the influence of concrete strength grade and boundary conditions on temperature and humidity diffusion, with the average error less than 5 %. The LOSS differences of traditional physics informed neural network (PINN) and BPINN in time step, activation function, hidden layer and neuron number are quantified. Compared with the traditional PINN, the LOSS of BPINN is reduced by 62.4 %. On this basis, the predictive performance of BPINN and four types of data-driven models is compared to verify the influence of physics constraint, as BPINN has the smallest statistical loss and data discreteness. The model proposed in this paper enhances the learning ability of spatial-temporal features by balancing the weight between boundary and physics equations, providing new insights for the thermo-hygro-mechanical coupling field in early-age concrete.

Set-to-set matching aims to identify correspondences between two sets of unordered items by minimizing a distance metric or maximizing a similarity measure. Traditional metrics, such as Chamfer Distance (CD) and Earth Mover’s Distance (EMD), are widely used for this purpose but often suffer from limitations like suboptimal performance in terms of accuracy and robustness, or high computational costs - or both. In this paper, we propose a novel, simple yet effective set-to-set matching similarity measure, GPS, based on Gumbel prior distributions. These distributions are typically used to model the extrema of samples drawn from various distributions. Our approach is motivated by the observation that the distributions of minimum distances from CD, as encountered in real world applications such as point cloud completion, can be accurately modeled using Gumbel distributions. We validate our method on tasks like few-shot image classification and 3D point cloud completion, demonstrating significant improvements over state of-the-art loss functions across several benchmark datasets. Our demo code is publicly available at https://github.com/Zhang-VISLab/ICLR2025-GPS

Iterative methods such as iterative closest point (ICP) for point cloud registration often suffer from bad local optimality (e.g. saddle points), due to the nature of nonconvex optimization. To address this fundamental challenge, in this paper we propose learning to form the loss landscape of a deep iterative method w.r.t. predictions at test time into a convex- like shape locally around each ground truth given data, namely Deep Loss Convexification (DLC), thanks to the overparametrization in neural networks. To this end, we formulate our learning objective based on adversarial training by manipulating the ground-truth predictions, rather than input data. In particular, we propose using star-convexity, a family of structured nonconvex functions that are unimodal on all lines that pass through a global minimizer, as our geometric constraint for reshaping loss landscapes, leading to (1) extra novel hinge losses appended to the original loss and (2) near-optimal predictions. We demonstrate the state-of-the-art performance using DLC with existing network architectures for the tasks of training recurrent neural networks (RNNs), 3D point cloud registration, and multimodel image alignment.

3D point clouds enhanced the robot’s ability to perceive the geometrical information of the environments, making it possible for many downstream tasks such as grasp pose detection and scene understanding. The performance of these tasks, though, heavily relies on the quality of data input, as incomplete can lead to poor results and failure cases. Recent training loss functions designed for deep learning-based point cloud completion, such as Chamfer distance (CD) and its variants (e.g. HyperCD [1]), imply a good gradient weighting scheme can significantly boost performance. However, these CD-based loss functions usually require data-related parameter tuning, which can be time-consuming for data-extensive tasks. To address this issue, we aim to find a family of weighted training losses (weighted CD) that requires no parameter tuning. To this end, we propose a search scheme, Loss Distillation via Gradient Matching, to find good candidate loss functions by mimicking the learning behavior in backpropagation between HyperCD and weighted CD. Once this is done, we propose a novel bilevel optimization formula to train the backbone network based on the weighted CD loss. We observe that: (1) with proper weighted functions, the weighted CD can always achieve similar performance to HyperCD, and (2) the Landau weighted CD, namely Landau CD, can outperform HyperCD for point cloud completion and lead to new state-of-the-art results on several benchmark datasets. Our demo code is available at https://github.com/Zhang-VISLab/IROS2024-LossDistillationWeightedCD.

In recent years, there has been a growing trend of incorporating hyperbolic geometry methods into computer vision. While these methods have achieved state-of-the-art performance on various metric learning tasks using hyperbolic distance measurements, the underlying theoretical analysis supporting this superior performance remains underexploited. In this study, we investigate the effects of integrating hyperbolic space into metric learning, particularly when training with contrastive loss. We identify a need for a comprehensive comparison between Euclidean and hyperbolic spaces regarding the temperature effect in the contrastive loss within the existing literature. To address this gap, we conduct an extensive investigation to benchmark the results of Vision Transformers (ViTs) using a hybrid objective function that combines loss from Euclidean and hyperbolic spaces. Additionally, we provide a theoretical analysis of the observed performance improvement. We also reveal that hyperbolic metric learning is highly related to hard negative sampling, providing insights for future work. This work will provide valuable data points and experience in understanding hyperbolic image embeddings. To shed more light on problem-solving and encourage further investigation into our approach, our code1is available online.

Chamfer distance (CD) is a standard metric to measure the shape dissimilarity between point clouds in point cloud completion, as well as a loss function for (deep) learning. However, it is well known that CD is vulnerable to outliers, leading to the drift towards suboptimal models. In contrast to the literature where most works address such issues in Euclidean space, we propose an extremely simple yet powerful metric for point cloud completion, namely Hyperbolic Chamfer Distance (HyperCD), that computes CD in hyperbolic space. In backpropagation, HyperCD consistently assigns higher weights to the matched point pairs with smaller Euclidean distances. In this way, good point matches are likely to be preserved while bad matches can be updated gradually, leading to better completion results. We demonstrate state-of-the-art performance on the benchmark datasets, i.e. PCN, ShapeNet-55, and ShapeNet34, and show from visualization that HyperCD can significantly improve the surface smoothness. Code is available at: https://github.com/Zhang-VISLab.

A point cloud is a discrete set of data points sampled from a 3D geometric surface. Chamfer distance (CD) is a popular metric and training loss to measure the distances between point clouds, but also well known to be sensitive to outliers. To address this issue, in this paper we propose InfoCD, a novel contrastive Chamfer distance loss to learn to spread the matched points for better distribution alignments between point clouds as well as accounting for a surface similarity estimator. We show that minimizing InfoCD is equivalent to maximizing a lower bound of the mutual information between the underlying geometric surfaces represented by the point clouds, leading to a regularized CD metric which is robust and computationally efficient for deep learning. We conduct comprehensive experiments for point cloud completion using InfoCD and observe significant improvements consistently over all the popular baseline networks trained with CD-based losses, leading to new state-of-the-art results on several benchmark datasets. Demo code is available at https://github.com/Zhang-VISLab/NeurIPS2023-InfoCD.

Point clouds measured by 3D scanning devices often have partially missing data due to the view positioning of the scanner. The missing data can reduce the performance of a point cloud in downstream tasks such as segmentation, location, and pose estimation. Consequently, 3D point cloud completion aims to predict the missing regions of incomplete objects for these fundamental 3D vision tasks. However, predicting the complete object can easily diminish the detail or structure of a measured region, which usually does not require repair. This study proposes a novel neural network architecture, Cosmos Propagation Network (CP-Net), for 3D point cloud completion. CP-Net extracts latent features in different scales from incomplete point clouds used as input. For point cloud generation, we propose a novel point expand method using a Mirror Expand module. Compared with existing methods, our Mirror Expand module introduces less information redundancy, which makes the distribution of points more reliable. CP-Net predicts the details of missing regions and maintains a clear general structure. The performance of CP-Net on several benchmarks was compared to that of current baseline methods. Compared to the existing methods, CP-Net showed the best performance for various metrics. Thus, CP-Net is expected to help address various problems related to 3D point cloud completion. Its source code is available at https://github.com/ark1234/CP-Net.

Air pollution is a societal and cross-boundary environmental problem that can be visualized using a satellite. Satellite imaging is not only useful to the home country but also to the neighboring countries. Moreover, monitoring the movement of air pollution can help susceptible people avoid acid rain and photochemical smog. Using advanced remote sensing (RS) images, substantial information can be obtained, which can produce numerous effective methods for visualizing air pollution. In this article, a novel method for extracting air pollution has been proposed; it applies various pipeline networks along with a focus area method to exploit the spectral aspect information. Afterward, three indices with numerous modified fully convolutional networks (FCNs) were extracted. Then, by employing a multivote module, visualized air pollution can be presented. In the conducted experiments, five-year Himawari-8 satellite images have been utilized in the North–East Asia area to validate the frameworks. Furthermore, the experimental result indicating that the given methods could effectively visualize air pollution. Source code and data sets are available athttps://github.com/ark1234/Himawari-8-based-visualized-understanding.

Text-based person search aims at retrieving the target person in an image gallery using a descriptive sentence of that person. The core of this task is to calculate a similarity score between the pedestrian image and description, which requires inferring the complex latent correspondence between image sub-regions and textual phrases at different scales. Transformer is an intuitive way to model the complex alignment by its self-attention mechanism. Most previous Transformer-based methods simply concatenate image region features and text features as input and learn a cross-modal representation in a brute force manner. Such weakly supervised learning approaches fail to explicitly build alignment between image region features and text features, causing an inferior feature distribution. In this paper, we present CFLT, Conditional Feature Learning based Transformer. It maps the sub-regions and phrases into a unified latent space and explicitly aligns them by constructing conditional embeddings where the feature of data from one modality is dynamically adjusted based on the data from the other modality. The output of our CFLT is a set of similarity scores for each sub-region or phrase rather than a cross-modal representation. Furthermore, we propose a simple and effective multi-modal re-ranking method named Re-ranking scheme by Visual Conditional Feature (RVCF). Benefit from the visual conditional feature and better feature distribution in our CFLT, the proposed RVCF achieves significant performance improvement. Experimental results show that our CFLT outperforms the state-of-the-art methods by 7.03% in terms of top-1 accuracy and 5.01% in terms of top-5 accuracy on the text-based person search dataset.