A Classroom Emotion Recognition Model Based on a Convolutional Neural Network Speech Emotion Algorithm

In this paper, we construct a model of convolutional neural network speech emotion algorithm, analyze the classroom identified by the neural network with a certain degree of confidence together with the school used in the dataset, find the characteristics and rules of teachers' control of classroom emotion nowadays using big data, find the parts of classroom emotion, and design a classroom emotion recognition model based on convolutional neural network speech emotion algorithm according to these characteristics. This paper will investigate the factors and patterns of teachers' emotional control in the classroom. In this paper, the existing neural network is adapted and improved, and some preprocessing is performed on the current dataset to train the network. The network used in this paper is a combination of convolutional neural network (CNN) and recurrent neural network (RNN), which takes advantage of both CNN for feature extraction and RNN for memory capability in the sequence model. This network has a good effect on both object labeling and speech recognition. For the problem of extracting emotion features of whole-sentence speech, we propose an attention mechanism-based emotion recognition algorithm for variable-length speech and design a spatiotemporal attention module for the speech emotion algorithm and a convolutional channel attention module for the CNN network to reduce the contribution of the spatiotemporal data of the speech emotion algorithm and the unimportant parts of the CNN convolutional channel feature data in the subsequent recognition by the attention mechanism. In turn, the weight of core key data and features is increased to improve the model recognition accuracy.

Improving Multi-Agent Generative Adversarial Nets with Variational Latent Representation

Generative adversarial networks (GANs), which are a promising type of deep generative network, have recently drawn considerable attention and made impressive progress. However, GAN models suffer from the well-known problem of mode collapse. This study focuses on this challenge and introduces a new model design, called the encoded multi-agent generative adversarial network (E-MGAN), which tackles the mode collapse problem by introducing the variational latent representations learned from a variable auto-encoder (VAE) to a multi-agent GAN. The variational latent representations are extracted from training data to replace the random noise input of the general multi-agent GANs. The generator in E-MGAN employs multiple generators and is penalized by a classifier. This integration guarantees that the proposed model not only enhances the quality of generated samples but also improves the diversity of generated samples to avoid the mode collapse problem. Moreover, extensive experiments are conducted on both a synthetic dataset and two large-scale real-world datasets. The generated samples are visualized for qualitative evaluation. The inception score (IS) and Fréchet inception distance (FID) are adopted to measure the performance of the model for quantitative assessment. The results confirmed that the proposed model achieves outstanding performances compared to other state-of-the-art GAN variants.