Learning exists in the context of data, yet notions of confidence typically focus on model predictions, not label quality. Confident learning (CL) is an alternative approach which focuses instead on ...label quality by characterizing and identifying label errors in datasets, based on the principles of pruning noisy data, counting with probabilistic thresholds to estimate noise, and ranking examples to train with confidence. Whereas numerous studies have developed these principles independently, here, we combine them, building on the assumption of a class-conditional noise process to directly estimate the joint distribution between noisy (given) labels and uncorrupted (unknown) labels. This results in a generalized CL which is provably consistent and experimentally performant. We present sufficient conditions where CL exactly finds label errors, and show CL performance exceeding seven recent competitive approaches for learning with noisy labels on the CIFAR dataset. Uniquely, the CL framework is not coupled to a specific data modality or model (e.g., we use CL to find several label errors in the presumed error-free MNIST dataset and improve sentiment classification on text data in Amazon Reviews). We also employ CL on ImageNet to quantify ontological class overlap (e.g., estimating 645 missile images are mislabeled as their parent class projectile), and moderately increase model accuracy (e.g., for ResNet) by cleaning data prior to training. These results are replicable using the open-source cleanlab release.
Organometallic halide perovskite films with good surface morphology and large grain size are desirable for obtaining high‐performance photovoltaic devices. However, defects and related trap sites are ...generated inevitably at grain boundaries and on surfaces of solution‐processed polycrystalline perovskite films. Seeking facial and efficient methods to passivate the perovskite film for minimizing defect density is necessary for further improving the photovoltaic performance. Here, a convenient strategy is developed to improve perovskite crystallization by incorporating a 2D polymeric material of graphitic carbon nitride (g‐C3N4) into the perovskite layer. The addition of g‐C3N4 results in improved crystalline quality of perovskite film with large grain size by retarding the crystallization rate, and reduced intrinsic defect density by passivating charge recombination centers around the grain boundaries. In addition, g‐C3N4 doping increases the film conductivity of perovskite layer, which is beneficial for charge transport in perovskite light‐absorption layer. Consequently, a champion device with a maximum power conversion efficiency of 19.49% is approached owing to a remarkable improvement in fill factor from 0.65 to 0.74. This finding demonstrates a simple method to passivate the perovskite film by controlling the crystallization and reducing the defect density.
Graphitic carbon nitride (g‐C3N4) is incorporated into the perovskite precursor solution to modify the perovskite film by controlling the perovskite crystallization, reducing the intrinsic defect density, and improving the film conductivity. As a result, a champion device with a maximum power conversion efficiency of 19.49% is approached.
Abstract
As a key system to ensure residential comfort, the control mode of the polar cruise air conditioning system directly relates to the comfort of the cabin environment and the economy of ship ...operation. Aiming at the characteristics of uncertainty and regulation time delay in the load change of the VAV air conditioning system in the public area of polar cruise ships, this paper models the mechanism of the VAV terminal system and conducts simulation experiments on the Matlab/Simulink simulation platform. By comparing the difference between the traditional PID control strategy and the fuzzy PID control strategy in the control effect, it provides a reference for the optimal control of the polar cruise air conditioning system.
Abstract
In this paper, a module fault identification method for a power battery system based on multi-classifier fusion is designed. Taking the inverter single tube open circuit fault of the power ...battery system as an example, the deep learning method of deep neural network and convolutional neural network is used to build the primary diagnosis layer. The fusion method of evidential reasoning rules is used to fuse the primary diagnosis results. The fault data are obtained through simulation of the simulation model, which verifies that the scheme avoids the disadvantage of unstable diagnosis results of a single classifier, and obtains more stable diagnostic results with higher diagnostic accuracy.
Abstract
As the ship is sailing in the sea, the PID control parameters set may not be the optimal parameters under different sea conditions according to the changes of sea conditions and wind, waves ...and currents, so the PID control of fin stabilizer cannot achieve the ideal control effect in the actual ship operation process. This paper studies the design of fuzzy controller with better robustness than traditional PID control, and explores the actual control performance of fuzzy controller under different sea conditions.
Abstract
This paper discussed the performance changes under a rated operation state and permanent magnet loss-of-field fault of a 5KW permanent magnet synchronous motor (PMSM) in a 200 mm shaftless ...rimless thruster. These changes in key parameters include air gap flux density, back electromotive force, output torque, current and magnetic field. Using the changes in these parameters to analyze the running state of the motor provides a reference for the early fault diagnosis of a permanent magnet synchronous motor at the simulation level.
The combination of scientific and technological achievements and sports has found new opportunities to change people’s sports habits. Sport training takes up an increasing proportion of people’s ...lives. In order to improve the efficiency of sports training and standardize the training actions of players, this article is based on wireless network communication and uses different types of recognition methods in the field of action recognition to build basic classifications. Iterative mutual training to improve generalization performance can reduce the cost of labeling and realize the complementary advantages of different recognition methods, thereby improving the recognition accuracy of human actions. Finally, the algorithm is used to recognize human movements. This method can effectively overcome the problem of differential degradation of base classifiers in the iterative process of collaborative training and further improve the accuracy of human action recognition. The experimental results prove that the motion recognition of wireless network communication proposed in this paper can effectively improve the accuracy of athletes’ movements, which is more than 20% higher than traditional methods, and, under the guidance of standardized movements, can reduce athletes’ sports injuries.
Display omitted
•Biochars are potential sustainable precursors for activated carbon production.•Physical activation and chemical activation are applied in the production process.•Production ...parameters affect the properties of resultant activated carbon.•Multiple applications in environmental protection and energy storage are reviewed.•Future perspectives about biochar activation and applications are highlighted.
There is a growing interest of the scientific community on production of activated carbon using biochar as potential sustainable precursors pyrolyzed from biomass wastes. Physical activation and chemical activation are the main methods applied in the activation process. These methods could have significantly beneficial effects on biochar chemical/physical properties, which make it suitable for multiple applications including water pollution treatment, CO2 capture, and energy storage. The feedstock with different compositions, pyrolysis conditions and activation parameters of biochar have significant influences on the properties of resultant activated carbon. Compared with traditional activated carbon, activated biochar appears to be a new potential cost-effective and environmentally-friendly carbon materials with great application prospect in many fields. This review not only summarizes information from the current analysis of activated biochar and their multiple applications for further optimization and understanding, but also offers new directions for development of activated biochar.
HPV vaccine can prevent HPV infection effectively. The college student’s vaccination status is unclear in mainland China. We assessed the knowledge, practice, and attitude towards HPV vaccine and ...compared the differences between medical and nonmedical students. It was a cross-sectional study using self-administered anonymous questionnaires. Nine-hundred sixty full-time college students were recruited randomly at Peking University in China. The medical students had higher level of knowledge of HPV and its vaccine than the nonmedical students (
p
< 0.001). The vaccinated female students were 9.0%. The high-grade clinical students had a higher uptake rate than the nonmedical students (19.5 vs 8.6%,
p
< 0.05). Awareness of HPV (
p
< 0.01), awareness of the vaccine (
p
< 0.001), and vaccinated family members or friends (
p
< 0.001) were related to the nonmedical students’ vaccination. Vaccinated family members or friends were significant predictor for students’ vaccination status (
p
< 0.001). Medical students knew more about HPV and its vaccine than nonmedical students. Female students’ vaccinated rate was low, and the high-grade clinical students had a higher uptake rate than the nonmedical students.
The pressure hull of deep manned submersible is the most basic component to ensure its intended function. It is necessary to study the influence of initial geometric defects on the bearing capacity ...of pressure hull of manned submersible with different depths. According to the idea of discontinuous Galerkin finite element method, the theoretical model is constructed and the corresponding algorithm is designed, and the genetic algorithm is combined with discontinuous Galerin finite element method to establish the inverse method to obtain the ultimate bearing capacity of manned submersible. First, the discontinuous Galerkin finite element model is constructed, the inversion model is also established through combing the discontinuous Galerkin finite element method and genetic algorithm, and then the corresponding solution algorithm is designed. Moreover, then, the ultimate bearing analysis of manned submersible for different deep is carried out based on the inversion model combing discontinuous Galerkin finite element method and genetic algorithm. The effect of defect parameters on ultimate bearing capacity of manned submersible is obtained.