Scalable batch-correction method for integrating large-scale single-cell transcriptomes

Integration of the evolving large-scale single-cell transcriptomes requires scalable batch-correction approaches. Here we propose a simple batch-correction method that is scalable for integrating super large-scale single-cell transcriptomes from diverse sources. The core idea of the method is encoding batch information of each cell as a trainable parameter and added to its expression profile; subsequently, a contrastive learning approach is used to learn feature representation of the additive expression profile. We demonstrate the scalability of the proposed method by integrating 18 million cells obtained from the Human Cell Atlas. Our benchmark comparisons with current state-of-the-art single-cell integration methods demonstrated that our method could achieve comparable data alignment and cluster preservation. Our study would facilitate the integration of super large-scale single-cell transcriptomes. The source code is available at https://github.com/xilinshen/Fugue.

Dynamic Intervisibility Analysis of 3D Point Clouds

With the popularity of ground and airborne three-dimensional laser scanning hardware and the development of advanced technologies for computer vision in geometrical measurement, intelligent processing of point clouds has become a hot issue in artificial intelligence. The intervisibility analysis in 3D space can use viewpoint, view distance, and elevation values and consider terrain occlusion to derive the intervisibility between two points. In this study, we first use the 3D point cloud of reflected signals from the intelligent autonomous driving vehicle’s 3D scanner to estimate the field-of-view of multi-dimensional data alignment. Then, the forced metrics of mechanical Riemann geometry are used to construct the Manifold Auxiliary Surface (MAS). With the help of the spectral analysis of the finite element topology structure constructed by the MAS, an innovative dynamic intervisibility calculation is finally realized under the geometric calculation conditions of the Mix-Planes Calculation Structure (MPCS). Different from advanced methods of global and interpolation pathway-based point clouds computing, we have removed the 99.54% high-noise background and reduced the computational complexity by 98.65%. Our computation time can reach an average processing time of 0.1044 s for one frame with a 25 fps acquisition rate of the original vision sensor. The remarkable experimental results and significant evaluations from multiple runs demonstrate that the proposed dynamic intervisibility analysis has high accuracy, strong robustness, and high efficiency. This technology can assist in terrain analysis, military guidance, and dynamic driving path planning, Simultaneous Localization And Mapping (SLAM), communication base station siting, etc., is of great significance in both theoretical technology and market applications.

A Self-Supervised Tibetan-Chinese Vocabulary Alignment Method

Tibetan is a low-resource language. In order to alleviate the shortage of parallel corpus between Tibetan and Chinese, this paper uses two monolingual corpora and a small number of seed dictionaries to learn the semi-supervised method with seed dictionaries and self-supervised adversarial training method through the similarity calculation of word clusters in different embedded spaces and puts forward an improved selfsupervised adversarial learning method of Tibetan and Chinese monolingual data alignment only. The experimental results are as follows. The seed dictionary of semi-supervised method made before 10 predicted word accuracy of 66.5 (Tibetan - Chinese) and 74.8 (Chinese - Tibetan) results, to improve the self-supervision methods in both language directions have reached 53.5 accuracy.

A Self-Supervised Tibetan-Chinese Vocabulary Alignment Method based on Adversarial Learning

Tibetan is a low-resource language. In order to alleviate the shortage of parallel corpus between Tibetan and Chinese, this paper uses two monolingual corpora and a small number of seed dictionaries to learn the semi-supervised method with seed dictionaries and self-supervised adversarial training method through the similarity calculation of word clusters in different embedded spaces and puts forward an improved self-supervised adversarial learning method of Tibetan and Chinese monolingual data alignment only. The experimental results are as follows. First, the experimental results of Tibetan syllables Chinese characters are not good, which reflects the weak semantic correlation between Tibetan syllables and Chinese characters; second, the seed dictionary of semi-supervised method made before 10 predicted word accuracy of 66.5 (Tibetan - Chinese) and 74.8 (Chinese - Tibetan) results, to improve the self-supervision methods in both language directions have reached 53.5 accuracy.

Assessing Place Type Similarities Based on Functional Signatures Extracted from Social Media Data

Place types are often used to query places or retrieve data in gazetteers. Existing gazetteers do not use the same place type classification schemes, and the various typing schemes can cause difficulties in data alignment and matching. Different place types may share some level of similarities. However, previous studies have paid little attention to the place type similarities. This study proposes an analytical approach to measuring similarities between place types in multiple typing schemes based on functional signatures extracted from web-harvested place descriptions. In this study, a functional signature consists of three component signature factors: place affordance, events, and key-descriptors. The proposed approach has been tested in a case study using Twitter data. The case study finds high similarity scores between some pairs of types and summarizes the situations when high similarities could occur. The research makes two innovative contributions: First, it proposes a new analytical approach to measuring place type similarities. Second, it demonstrates the potential and benefits of using location-based social media data to better understand places.

A Novel Approach to Global Positioning System Accuracy Assessment, Verified on LiDAR Alignment of One Million Kilometers at a Continent Scale, as a Foundation for Autonomous Driving Safety Analysis

This paper concerns a new methodology for accuracy assessment of GPS (Global Positioning System) verified experimentally with LiDAR (Light Detection and Ranging) data alignment at continent scale for autonomous driving safety analysis. Accuracy of an autonomous driving vehicle positioning within a lane on the road is one of the key safety considerations and the main focus of this paper. The accuracy of GPS positioning is checked by comparing it with mobile mapping tracks in the recorded high-definition source. The aim of the comparison is to see if the GPS positioning remains accurate up to the dimensions of the lane where the vehicle is driving. The goal is to align all the available LiDAR car trajectories to confirm the of accuracy of GNSS+INS (Global Navigation Satellite System + Inertial Navigation System). For this reason, the use of LiDAR metric measurements for data alignment implemented using SLAM (Simultaneous Localization and Mapping) was investigated, assuring no systematic drift by applying GNSS+INS constraints. The methodology was verified experimentally using arbitrarily chosen measurement instruments (NovAtel GNSS+INS, Velodyne HDL32 LiDAR) mounted onto mobile mapping systems. The accuracy was assessed and confirmed by the alignment of 32,785 trajectories with a total length of 1,159,956.9 km and a total of 186.4 × 109 optimized parameters (six degrees of freedom of poses) that cover the United States region in the 2016–2019 period. The alignment improves the trajectories; thus the final map is consistent. The proposed methodology extends the existing methods of global positioning system accuracy assessment, focusing on realistic environmental and driving conditions. The impact of global positioning system accuracy on autonomous car safety is discussed. It is shown that 99% of the assessed data satisfy the safety requirements (driving within lanes of 3.6 m) for Mid-Size (width 1.85 m, length 4.87 m) vehicles and 95% for Six-Wheel Pickup (width 2.03–2.43 m, length 5.32–6.76 m). The conclusion is that this methodology has great potential for global positioning accuracy assessment at the global scale for autonomous driving applications. LiDAR data alignment is introduced as a novel approach to GNSS+INS accuracy confirmation. Further research is needed to solve the identified challenges.