3D Gaussian Splatting-based Open-vocabulary 3D segmentation has shown impressive performance with dense input images. However, existing methods exhibit poor results when confronted with sparse inputs, primarily due to limited overlap among input views and insufficient view supervision provided. To tackle these challenges, we propose SpContrast, a novel framework that creates additional semantic constraints to enhance sparse-view 3D open-vocabulary segmentation. First, we introduce Collaborative Contrastive Learning (CCL), which creates instructive multi-view semantic constraints by collaboratively mining semantic interactions between training and online-rendered novel views. Motivated by the principle that semantically consistent features should converge and divergent ones separate, CCL establishes cross-view contrastive constraints to enhance semantic coherence. Second, to alleviate the adverse impact of false negative samples caused by semantic inconsistencies within the same object, we present Region-aware Negative Sampling (RNS). RNS rectifies these false negative samples, and treats them as hard samples during our contrastive optimization, leading to improved object completeness and more accurate segmentation. Extensive experiments on challenging sparse-input datasets, including Replica and ScanNet, demonstrate the superiority of SpContrast, achieving 7.6% and 8.3% mIoU improvements for 3D open-vocabulary segmentation.

The outbreak of COVID-19 has resulted in an "infodemic" that has encouraged the propagation of misinformation about COVID-19 and cure methods which, in turn, could negatively affect the adoption of recommended public health measures in the larger population. In this paper, we provide a new multimodal (consisting of images, text and temporal information) labeled dataset containing news articles and tweets on the COVID-19 vaccine. We collected 2,593 news articles from 80 publishers for one year between Feb 16th 2020 to May 8th 2021 and 24184 Twitter posts (collected between April 17th 2021 to May 8th 2021). We combine ratings from two news media ranking sites: Medias Bias Chart and Media Bias/Fact Check (MBFC) to classify the news dataset into two levels of credibility: reliable and unreliable. The combination of two filters allows for higher precision of labeling. We also propose a stance detection mechanism to annotate tweets into three levels of credibility: reliable, unreliable and inconclusive. We provide several statistics as well as other analytics like, publisher distribution, publication date distribution, topic analysis, etc. We also provide a novel architecture that classifies the news data into misinformation or truth to provide a baseline performance for this dataset. We find that the proposed architecture has an F-Score of 0.919 and accuracy of 0.882 for fake news detection. Furthermore, we provide benchmark performance for misinformation detection on tweet dataset. This new multimodal dataset can be used in research on COVID-19 vaccine, including misinformation detection, influence of fake COVID-19 vaccine information, etc.

Monsoon coastal cities often suffer from extreme rain-induced flooding and severe hazard. However, the associated physical mechanisms and detailed storm structures are poorly understood due to the lack of high-resolution data. This study presents an analysis of a thunderstorm that produces extreme hourly rainfall (EXHR) of 219 mm over the Guangzhou megacity on the southern coast of China using integrated multiplatform observations and a four-dimensional variational Doppler radar analysis system. Results indicate that weak environmental flows and convectively generated weak cold pool facilitate the formation of a quasi-stationary storm, while onshore warm and moist flows in the boundary layer (BL) provide the needed moisture supply. The 219-mm EXHR is attendant by a shallow meso-$\gamma $-scale vortex due to stretching of intense latent heating-induced convergence, which, in turn, helps organize convective updrafts into its core region. Lightning and dual-polarization radar observations reveal active warm-rain (but weak mixed-phase) microphysical processes, with raindrop size distribution (RSD) closer to marine convection. In contrast, another storm develops about 4 h earlier and only 35 km to the northwest, but with more lightning, higher cloud tops, more graupel and supercooled liquid water content, more continental RSD, little evidence of rotation, and much less rainfall; they are attributable to the presence of larger convective available potential energy resulting from the urban heat island effects and less moisture supply in the BL. These results highlight the importance of using multisource remote sensing data sets in understanding the microphysical and kinematic structures of EXHR-producing storms.

A large number of robots have put forward the new requirements for humanrobot interaction. One of the problems in human-swarm robot interaction is how to naturally achieve an efficient and accurate interaction between humans and swarm robot systems. To address this, this paper proposes a new type of human-swarm natural interaction system. Through the cooperation between three-dimensional (3D) gesture interaction channel and natural language instruction channel, a natural and efficient interaction between a human and swarm robots is achieved. First, A 3D lasso technology realizes a batch-picking interaction of swarm robots through oriented bounding boxes. Second, control instruction labels for swarm-oriented robots are defined. The instruction label is integrated with the 3D gesture and natural language through instruction label filling. Finally, the understanding of natural language instructions is realized through a text classifier based on the maximum entropy model. A head-mounted augmented reality display device is used as a visual feedback channel. The experiments on selecting robots verify the feasibility and availability of the system.

With the rapidly growing number of connected smart devices deployed and diverse services provided in the Internet of Things (IoT), selecting proper services for users is becoming more and more important. However, challenges exist as a result of the highly heterogeneous environments, characteristics of various kinds of users and the myriad services offered by many service providers, which have promising applications in the IoT era. In the meantime, users' contexts (e.g., location, time, and surroundings) are wildly utilized in the IoT scenario to better satisfy individuals' demands, raising privacy issues among people. To address these problems, we proposed a differentially private tree-based contextual online learning approach for IoT service selection to select suitable services for users. Leveraging on the historical records of services' and users' feedback, our algorithm achieves high prediction accuracy. Besides, instead of considering the services as individual items, we utilize a top-down cover tree structure to select services, which supports increasing large-scale dataset and diverse natural conditions. We theoretically prove that the accumulative regret of our approach has a sublinear bound and our experiment confirms that it can handle big data problems while achieving a balance between privacy-preserving level and service selection accuracy.