Positive and Unlabeled (PU) learning refers to a special case of binary classification, and technically, it aims to induce a binary classifier from a few labeled positive training instances and loads of unlabeled inst...
Positive and Unlabeled (PU) learning refers to a special case of binary classification, and technically, it aims to induce a binary classifier from a few labeled positive training instances and loads of unlabeled instances. In this paper, we derive a theorem indicating that the probability boundary of the asymmetric disambiguation-free expected risk of PU learning is controlled by its asymmetric penalty, and we further empirically evaluated this theorem. Inspired by the theorem and its empirical evaluations, we propose an easy-to-implement two-stage PU learning method, namely Positive and Unlabeled Learning with Controlled Probability Boundary Fence (PUL-CPBF). In the first stage, we train a set of weak binary classifiers concerning different probability boundaries by minimizing the asymmetric disambiguation-free empirical risks with specific asymmetric penalty values. We can interpret these induced weak binary classifiers as a probability boundary fence. For each unlabeled instance, we can use the predictions to locate its class posterior probability and generate a stochastic label. In the second stage, we train a strong binary classifier over labeled positive training instances and all unlabeled instances with stochastic labels in a self-training manner. Extensive empirical results demonstrate that PUL-CPBF can achieve competitive performance compared with the existing PU learning baselines.
Visual odometry is one of the key core technologies in the field of autonomous driving. However, images captured in low-light or unevenly-illuminated scenes still cannot guarantee good performance due to low image con...
详细信息
Gene regulatory network(GRN) describes dynamic and complex gene interactions that determine the functions of ***,understanding of GRNs has played a great role in cancer treatment,drug design,and gene ***,the GRN study...
详细信息
Gene regulatory network(GRN) describes dynamic and complex gene interactions that determine the functions of ***,understanding of GRNs has played a great role in cancer treatment,drug design,and gene ***,the GRN study suffers from a lack of experiment measurement data and the complexity of the *** approaches have evolved in recent years to promote the study of *** this review,we summarized popular computational approaches for GRN *** started with traditional bioinformatic methods,such as Bayesian networks and mutual information ***,we introduced how today's hot technology in the computer field-machine learning benefited GRN ***-based approaches and other machine-learning methods are elaborated on in this *** discussed not only the advantages and progression brought by various methods but also the drawbacks and limitations,such as the accuracy and robustness of GRN *** expect to inspire readers for improved GRN study approaches via our introduction to this field.
In computer image processing, optical flow is a classic task used to track the motion of pixels. In the field of robotics, methods like SLAM extensively employ sparse optical flow as a substitute for time-consuming fe...
In computer image processing, optical flow is a classic task used to track the motion of pixels. In the field of robotics, methods like SLAM extensively employ sparse optical flow as a substitute for time-consuming feature point matching methods, enabling rapid and stable camera tracking. However, the widespread use of fisheye cameras introduces distortion issues in practical applications. Traditional sparse optical flow struggles to accurately track pixel positions in such scenarios, affecting the precision of camera pose tracking. Therefore, we introduce a sparse optical flow method designed for fisheye camera models, which correctly utilizes prior information obtained during camera tracking to enhance the accuracy of sparse optical flow. Finally, we propose a scheme for constructing a semi-distortion lookup table to accelerate distortion operations and reduce memory requirements. We conducted extensive testing on simulated datasets with various camera motion scenarios, and the experimental results demonstrate that our method outperforms OpenCV in terms of accuracy and versatility.
In this paper, we investigate a transmission eigenvalue problem that couples the principles of acoustics and elasticity. This problem naturally arises when studying fluid-solid interactions and constructing bubbly-ela...
详细信息
Most of the current visual odometry (VO) methods are designed based on a set of standard procedures, including camera calibration, feature extraction, feature matching (or tracking), motion estimation, local optimizat...
详细信息
The aim is to construct a country-dimension knowledge graph of COVID-19 vaccines from the information of COVID-19 vaccines and to analyze the leading countries of vaccine R&D by combining the advantages of easy op...
The aim is to construct a country-dimension knowledge graph of COVID-19 vaccines from the information of COVID-19 vaccines and to analyze the leading countries of vaccine R&D by combining the advantages of easy operation and intuitive feeling of knowledge graph visualization,to provide a reference for Chinese vaccine R&D departments and international *** this paper,through data collection,based on entity extraction and relationship construction,a knowledge graph of country dimensions was established by specifying the central vaccine R&D countries and vaccine distribution,and multidimensional microdata such as word frequency and betweenness centrality were combined to analyze the national characteristics of the COVID-19 *** analysis of the knowledge graph of the country dimension of the COVID-19 vaccine shows that countries with robust technology and economies,such as the US and China,choose to develop vaccine distribution independently,countries with advanced economies,such as Saudi Arabia,decide to purchase vaccine distribution,and less developed countries,such as South Africa and Latin America,need international aid for vaccines or purchase low-cost *** paper constructs the correlation between nodes and nodes of the COVID-19 vaccine with the help of a knowledge graph,systematically and comprehensively reveals the research mainstay and distribution model of the COVID-19 vaccine from the national level,and provides rationalized suggestions for international cooperation in vaccine R&D in China.
Human skeleton-based action recognition has long been an indispensable aspect of artificial intelligence. Current state-of-the-art methods tend to consider only the dependencies between connected skeletal joints, limi...
详细信息
Visual perception plays an important role in autonomous driving technology. The two key factors in visual perception tasks are monocular object detection and structured data analysis. In this paper, a structured data ...
详细信息
This paper focuses on the optimization and improvement of the line-of-sight tracking algorithm based on monocular vision, and aims to achieve a series of complex functions through image analysis based on line-of-sight...
详细信息
暂无评论