Search

Wheat is a critical crop, extensively consumed worldwide, and its production enhancement is essential to meet escalating demand. The presence of diseases like stem rust, leaf rust, yellow rust, and tan spot significantly diminishes wheat yield, making the early and precise identification of these diseases vital for effective disease management. With advancements in deep learning algorithms, researchers have proposed many methods for the automated detection of disease pathogens; however, accurately detectingmultiple disease pathogens simultaneously remains a challenge. This challenge arises due to the scarcity of RGB images for multiple diseases, class imbalance in existing public datasets, and the difficulty in extracting features that discriminate between multiple classes of disease pathogens. In this research, a novel method is proposed based on Transfer Generative Adversarial Networks for augmenting existing data, thereby overcoming the problems of class imbalance and data scarcity. This study proposes a customized architecture of Vision Transformers (ViT), where the feature vector is obtained by concatenating features extracted from the custom ViT and Graph Neural Networks. This paper also proposes a Model Agnostic Meta Learning (MAML) based ensemble classifier for accurate classification. The proposedmodel, validated on public datasets for wheat disease pathogen classification, achieved a test accuracy of 99.20% and an F1-score of 97.95%. Compared with existing state-of-the-art methods, this proposed model outperforms in terms of accuracy, F1-score, and the number of disease pathogens detection. In future, more diseases can be included for detection along with some other modalities like pests and weed. [ABSTRACT FROM AUTHOR]

Aiming at the problem that the vibration signal of planetary gearboxes has strong non-stationary characteristics， few fault samples and the dependence of traditional deep learning on data， an intelligent diagnosis method for planetary gearboxes based on continuous wavelet transform（CWT） and model agnostic meta learning（MAML） is proposed. First， the vibration signal of the planetary gearbox is converted into a time-frequency image through CWT， which effectively expresses the non-stationary characteristics of the planetary gearbox； then， the ability of “learning to learn” of MAML is used to train small samples of time-frequency images， and finally the “unseen” faults of planetary gearboxes are tested. Through fault diagnosis experiments of planetary gearboxes under different sample sizes， working conditions and noise environments， a conclusion is drawn that the proposed method has higher recognition accuracy， generalization and robustness compared with other methods.

Automated deep learning is a new development direction of deep learning based on automated machine learning. Its main purpose is to achieve the automated operation of deep learning algorithm design process from design, training, parameter adjustment and so on. Due to a large number of repeated operations such as parameter adjustment and optimization model, the algorithm design process relies heavily on designers and consumes a lot of design time. Based on the automated deep learning method, this paper designs a short term prediction method of thermal hydraulic transient operation parameters, which can achieve the whole design process hosting operation. This method only needs the designer to upload the data set to the prediction model to complete the algorithm design, and does not need the designer's intervention in the whole process, so as to simplify the operation and save the design time. At the same time, quantum genetic algorithm and model agnostic meta learning are introduced to ensure that the algorithm is still reliable and effective when the data set is expanded and the prediction problem is more complex. And the algorithm can be porting and application on quantum computer. In order to verify the algorithm, the steam mass flow rate at the outlet of main steam pipe of steam generator and the water temperature at the bottom of pressurizer are predicted respectively after 1 s, 2 s, 3 s and 5 s. The calculations show that the maximum error of the prediction models is about 4%, and the average prediction time of a group of parameters is about 0.7 ms. The prediction model based on the combination of the above algorithms can be used to predict the change trend of thermal transient operation parameters in a short time and provide effective judgment basis for the reactor accident early warning. [ABSTRACT FROM AUTHOR]