Speech emotion recognition is a substantial component of natural language processing (NLP). It has strict requirements for the effectiveness of feature extraction and that of the acoustic model. With that in mind, a Heterogeneous Parallel Convolution Bi-LSTM model is proposed to address the challenges. It consists of two heterogeneous branches: the left one contains two dense layers and a Bi-LSTM layer, while the right one contains a dense layer, a convolution layer, and a Bi-LSTM layer. It can exploit the spatiotemporal information more effectively, and achieves 84.65%, 79.67%, and 56.50% unweighted average recalls on the benchmark databases EMODB, CASIA, and SAVEE, respectively. Compared with the previous research results, the proposed model achieves better performance stably.

To mitigate the threat to power system caused by ramp events – large wind power fluctuation, this study proposes an advanced ramp prediction approach based on event detection framework. This approach contains two successive stages of work, including wind power forecasting and ramp detection. Considering high‐performance ramp prediction requires long‐term and accurate wind power prediction results; this study also proposes a hybrid prediction model at the first stage. By using wind power curve to reflect the physic mechanism of wind power generation, data from numerical weather prediction system could be used to realise long‐term trend prediction. Then, a multivariate model is built with a data‐mining algorithm to correct system errors of the primary prediction, which is addressed to improve long‐term prediction performance. At the second stage, a modified swinging door algorithm is applied for ramp detection. Performance of both the proposed long‐term wind power prediction and the corresponding ramp prediction are computed and compared with conventional models on an actual wind dataset. Comprehensive results validated the feasibility and superiority of the proposed ramp prediction approach.

Accurate tomato maturity detection represents a critical challenge in precision agriculture. A YOLOv11-based algorithm named YOLO-PGC is proposed in this study for tomato maturity detection. Its three innovative components are denoted by “PGC”, respectively representing the Polarization State Space Strategy with Dynamic Weight Allocation, the Global Horizontal–Vertical Context Module, and the Convolutional–Inductive Feature Fusion Module. The Polarization Strategy enhances robustness against occlusion through adaptive feature importance modulation, he Global Context Module integrates cross-dimensional attention mechanisms with hierarchical feature extraction, and the Convolutional–Inductive Feature Fusion Module employs multimodal integration for improved object discrimination in complex scenes. Experimental results demonstrate that YOLO-PGC achieves superior precision and mean average precision compared to state-of-the-art methods. Validation on the COCO benchmark confirms the framework’s generalization capabilities, maintaining computational efficiency for real-time deployment. YOLO-PGC establishes new performance standards for agricultural object detection with potential applications in similar computer vision challenges. Overall, these components and strategies are integrated into YOLO-PGC to achieve robust object detection in complex scenarios.