The statistical pyramid dense time-delay neural network (SPD-TDNN) model makes it difficult to deal with the imbalance of training data, poses a high risk of overfitting, and has weak generalization ability. To solve these problems, we propose a method based on the joint loss function and improved statistical pyramid dense time-delay neural network (JLF-ISPD-TDNN), which improves on the SPD-TDNN model and uses the joint loss function method to combine the advantages of the cross-entropy loss function and the comparative learning of the loss function. By minimizing the distance between speech embeddings from the same speaker and maximizing the distance between speech embeddings from different speakers, the model could achieve enhanced generalization performance and more robust speaker feature representation. We evaluated the proposed method’s performance using the evaluation indexes of the equal error rate (EER) and minimum cost function (minDCF). The experimental results show that the EEE and minDCF on the Aishell-1 dataset reached 1.02% and 0.1221%, respectively. Therefore, using the joint loss function in the improved SPD-TDNN model can significantly enhance the model’s speaker recognition performance.

In deep learning-based speaker verification, the loss function plays a crucial role. Most systems rely on a single loss function, or simply sum multiple losses with manually adjusted weights, increasing experimental complexity and failing to fully leverage the complementary characteristics of different losses. To address this, this paper proposes a speaker verification system based on channel attention and adaptive joint loss optimization. An adaptive joint loss function dynamically adjusts loss weights, allowing the model to better learn the similarities and differences of speakers, narrowing the gap between closed- and open-set testing, and enhancing generalization ability. A channel attention squeeze-and-excitation module is designed to improve the network’s ability to extract channel-specific features. On the AISHELL-1 dataset, the system achieved an equal error rate of 0.84% and a minimum detection cost function of 0.0528. Experimental results demonstrate a significant improvement in speaker verification performance, confirming the effectiveness of the proposed system.

To enhance speech emotion recognition capability, this study constructs a speech emotion recognition model integrating the adaptive acoustic mixup (AAM) and improved coordinate and shuffle attention (ICASA) methods. The AAM method optimizes data augmentation by combining a sample selection strategy and dynamic interpolation coefficients, thus enabling information fusion of speech data with different emotions at the acoustic level. The ICASA method enhances feature extraction capability through dynamic fusion of the improved coordinate attention (ICA) and shuffle attention (SA) techniques. The ICA technique reduces computational overhead by employing depth-separable convolution and an h-swish activation function and captures long-range dependencies of multi-scale time-frequency features using the attention weights. The SA technique promotes feature interaction through channel shuffling, which helps the model learn richer and more discriminative emotional features. Experimental results demonstrate that, compared to the baseline model, the proposed model improves the weighted accuracy by 5.42% and 4.54%, and the unweighted accuracy by 3.37% and 3.85% on the IEMOCAP and RAVDESS datasets, respectively. These improvements were confirmed to be statistically significant by independent samples t-tests, further supporting the practical reliability and applicability of the proposed model in real-world emotion-aware speech systems.