Perception plays an important role in reliable decision-making for autonomous vehicles. Over the last ten years, huge advances have been made in the field of perception. However, perception in extreme weather conditions is still a difficult problem,
especially in rainy weather conditions. In order to improve the detection effect of road targets in rainy environments, we analyze the physical characteristics of the rain layer and propose a deraining convolutional neural network structure. Based on this network structure, we design an ablation experiment and experiment results show that our method can effectively improve the accuracy of object detection in rainy conditions.
Parental control apps, which are mobile apps that allow parents to monitor and restrict their childrens activities online, are becoming increasingly adopted by parents as a means of safeguarding their childrens online safety. However, it is not clear
whether these apps are always beneficial or effective in what they aim to do; for instance, the overuse of restriction and surveillance has been found to undermine parent-child relationship and childrens sense of autonomy. In this work, we investigate this gap, asking specifically: how might childrens and parents perceptions be related to how parental control features were designed? To investigate this question, we conducted an analysis of 58 top Android parental control apps designed for the purpose of promoting childrens online safety, finding three major axes of variation in how key restriction and monitoring features were realised: granularity, feedback/transparency, and parent-child communications support. To relate these axes to perceived benefits and problems, we then analysed 3264 app reviews to identify references to aspects of the each of the axes above, to understand childrens and parents views of how such dimensions related to their experiences with these apps. Our findings led towards 1) an understanding of how parental control apps realise their functionalities differently along three axes of variation, 2) an analysis of exactly the ways that such variation influences childrens and parents perceptions, respectively of the usefulness or effectiveness of these apps, and finally 3) an identification of design recommendations and opportunities for future apps by contextualising our findings within existing digital parenting theories.
Given multiple source domains, domain generalization aims at learning a universal model that performs well on any unseen but related target domain. In this work, we focus on the domain generalization scenario where domain shifts occur among class-con
ditional distributions of different domains. Existing approaches are not sufficiently robust when the variation of conditional distributions given the same class is large. In this work, we extend the concept of distributional robust optimization to solve the class-conditional domain generalization problem. Our approach optimizes the worst-case performance of a classifier over class-conditional distributions within a Wasserstein ball centered around the barycenter of the source conditional distributions. We also propose an iterative algorithm for learning the optimal radius of the Wasserstein balls automatically. Experiments show that the proposed framework has better performance on unseen target domain than approaches without domain generalization.
Superconductivity was recently discovered in rhombohedral trilayer graphene (RTG) in the absence of a moire potential. Intringuigly, superconductivity is observed proximate to a metallic state with reduced isospin symmetry, but it remains unknown whe
ther this is a coincidence or a key ingredient for superconductivity. Using a Hartree-Fock analysis and constraints from experiments, we argue that the symmetry breaking is inter-valley coherent (IVC) in nature. We evaluate IVC fluctuations as a possible pairing glue, and find that they lead to unconventional superconductivity which is $p$-wave when fluctuations are strong. We further elucidate how the inter-valley Hunds coupling determines the spin-structure of the IVC ground state and breaks the degeneracy between spin-singlet and triplet superconductivity. Intriguingly, if the normal state is spin-unpolarized, we find that a ferromagnetic Hunds coupling favors spin-singlet superconductivity, in agreement with experiments. Instead, if the normal state is spin-polarized, then IVC fluctuations lead to spin-triplet pairing.
Accurate forecasting of traffic conditions is critical for improving safety, stability, and efficiency of a city transportation system. In reality, it is challenging to produce accurate traffic forecasts due to the complex and dynamic spatiotemporal
correlations. Most existing works only consider partial characteristics and features of traffic data, and result in unsatisfactory performances on modeling and forecasting. In this paper, we propose a periodic spatial-temporal deep neural network (PSTN) with three pivotal modules to improve the forecasting performance of traffic conditions through a novel integration of three types of information. First, the historical traffic information is folded and fed into a module consisting of a graph convolutional network and a temporal convolutional network. Second, the recent traffic information together with the historical output passes through the second module consisting of a graph convolutional network and a gated recurrent unit framework. Finally, a multi-layer perceptron is applied to process the auxiliary road attributes and output the final predictions. Experimental results on two publicly accessible real-world urban traffic data sets show that the proposed PSTN outperforms the state-of-the-art benchmarks by significant margins for short-term traffic conditions forecasting
Automatic detection of rail track and its fasteners via using continuously collected railway images is important to maintenance as it can significantly improve maintenance efficiency and better ensure system safety. Dominant computer vision-based det
ection models typically rely on convolutional neural networks that utilize local image features and cumbersome prior settings to generate candidate boxes. In this paper, we propose a deep convolutional transformer network based method to detect multi-class rail components including the rail, clip, and bolt. We effectively synergize advantages of the convolutional structure on extracting latent features from raw images as well as advantages of transformers on selectively determining valuable latent features to achieve an efficient and accurate performance on rail component detections. Our proposed method simplifies the detection pipeline by eliminating the need of prior settings, such as anchor box, aspect ratio, default coordinates, and post-processing, such as the threshold for non-maximum suppression; as well as allows users to trade off the quality and complexity of the detector with limited training data. Results of a comprehensive computational study show that our proposed method outperforms a set of existing state-of-art approaches with large margins
Automated inspection and detection of foreign objects on railways is important for rail transportation safety as it helps prevent potential accidents and trains derailment. Most existing vision-based approaches focus on the detection of frontal intru
sion objects with prior labels, such as categories and locations of the objects. In reality, foreign objects with unknown categories can appear anytime on railway tracks. In this paper, we develop a semi-supervised convolutional autoencoder based framework that only requires railway track images without prior knowledge on the foreign objects in the training process. It consists of three different modules, a bottleneck feature generator as encoder, a photographic image generator as decoder, and a reconstruction discriminator developed via adversarial learning. In the proposed framework, the problem of detecting the presence, location, and shape of foreign objects is addressed by comparing the input and reconstructed images as well as setting thresholds based on reconstruction errors. The proposed method is evaluated through comprehensive studies under different performance criteria. The results show that the proposed method outperforms some well-known benchmarking methods. The proposed framework is useful for data analytics via the train Internet-of-Things (IoT) systems
It has become a trend to use study with me (SWM) Livestream to create a personalized study ambiance. However, we still have little understanding of the activities of SWM livestream and the streamers motivation to produce SWM livestream. This paper pr
ovides an overview of the activities and how streamers regulate these activities of SWM livestream on a Chinese popular User Generated Content(UGC) website, Bilibili. We observed the number and popularity of the SWM livestreams and analyzed 800 livestreams to understand the streamers study goals. We analyzed 20 SWM livestreams in detail and interviewed 12 streamers and 10 viewers to understand the activities and the streamers motivation. We found that streamers produced SWM livestream to seek supervision, find like-minded study partners and help and company others. Streamers dont interact or instruct with the viewers directly but use chat-bot and autonomous interaction to alleviated the interaction burden. Unique sessions like checking-in and study progress reporting promote the viewers social presence, promoting SOC, and enhancing their engagement. Strict rules and punishment are widely used to concentrate the members on study and contribute to positive atmosphere. We also found that SWM livestream often disappears when the examination is done and the streamer faces doubts on motivation and appearance. These findings suggest that SRL community can provide cognitive and socioemotional support for lonely learners to stick to a long-term study. The activities and streamers practice inspired how streamers can focus on contemplative efforts while controlling the interaction.
X-ray observations provide a unique probe of the accretion disk corona of supermassive black holes (SMBHs). In this paper, we present a uniform emph{Chandra} X-ray data analysis of a sample of 152 $zgeq4.5$ quasars. We firmly detect 46 quasars of thi
s sample in 0.5-2~keV above 3~$sigma$ and calculate the upper limits of the X-ray flux of the remaining. We also estimate the power law photon index of the X-ray spectrum of 31 quasars. 24 of our sample quasars are detected in the FIRST or NVSS radio surveys; all of them are radio-loud. We statistically compare the X-ray properties of our $zgeq4.5$ quasars to other X-ray samples of AGN at different redshifts. The relation between the rest-frame X-ray luminosity and other quasar parameters, such as the bolometric luminosity, UV luminosity, or SMBH mass, show large scatters. These large scatters can be attributed to the narrow luminosity range at the highest redshift, the large measurement error based on relatively poor X-ray data, and the inclusion of radio-loud quasars in the sample. The $L_{rm X}-L_{rm UV}$ relationship is significantly sub-linear. We do not find a significant redshift evolution of the $L_{rm X}-L_{rm UV}$ relation, expressed either in the slope of this relation, or the departure of individual AGNs from the best-fit $alpha_{rm OX}-L_{rm UV}$ relation ($Deltaalpha_{rm OX}$). The median value of the X-ray photon index is $Gammaapprox1.79$, which does not show redshift evolution from $z=0$ to $zsim7$. The X-ray and UV properties of the most distant quasars could potentially be used as a standard candle to constrain cosmological models. The large scatter of our sample on the Hubble diagram highlights the importance of future large unbiased deep X-ray and radio surveys in using quasars in cosmological studies.
This paper presents SPICE, a Semantic Pseudo-labeling framework for Image ClustEring. Instead of using indirect loss functions required by the recently proposed methods, SPICE generates pseudo-labels via self-learning and directly uses the pseudo-lab
el-based classification loss to train a deep clustering network. The basic idea of SPICE is to synergize the discrepancy among semantic clusters, the similarity among instance samples, and the semantic consistency of local samples in an embedding space to optimize the clustering network in a semantically-driven paradigm. Specifically, a semantic-similarity-based pseudo-labeling algorithm is first proposed to train a clustering network through unsupervised representation learning. Given the initial clustering results, a local semantic consistency principle is used to select a set of reliably labeled samples, and a semi-pseudo-labeling algorithm is adapted for performance boosting. Extensive experiments demonstrate that SPICE clearly outperforms the state-of-the-art methods on six common benchmark datasets including STL10, Cifar10, Cifar100-20, ImageNet-10, ImageNet-Dog, and Tiny-ImageNet. On average, our SPICE method improves the current best results by about 10% in terms of adjusted rand index, normalized mutual information, and clustering accuracy.
