Develop A Hybrid Deep Learning Model Combining Spiking Neural Networks (SNN) Neural Ordinary Differential

Develop a hybrid deep learning model combining Spiking Neural Networks (SNN), Neural Ordinary Differential Equations (Neural ODE), and Gated Recurrent Units (GRU) for robust emotion recognition from speech. Leverage SNN to efficiently capture temporal dynamics of speech signals with low energy consumption. Utilize Neural ODE for continuous-time modeling of speech features, enabling smooth representation of emotional variations. Incorporate GRU to learn sequential dependencies and contextual patterns in speech for improved classification accuracy. Evaluate the model’s performance across benchmark emotional speech datasets to validate efficiency, scalability, and generalization

02.09.2025 14:46