The recent advancements in the field of edge computing and neural networks are significantly pushing the boundaries of efficiency and performance. Researchers are focusing on optimizing spiking neural networks (SNNs) for better accuracy and energy efficiency, particularly in resource-constrained environments. Novel encoding methods and partitioning strategies are being developed to enhance the computational capabilities of SNNs, enabling them to handle complex tasks with reduced energy consumption. Additionally, there is a growing emphasis on model pruning and compression techniques to deploy multi-object tracking and deep neural networks on edge devices without compromising accuracy. The integration of representation similarity in model merging schemes is also emerging as a promising direction, offering a more efficient way to manage memory and computational resources. Furthermore, the development of small language models optimized for edge AI is addressing the need for real-time NLP applications in low-resource settings. These innovations collectively aim to bridge the gap between the computational demands of modern AI models and the limitations of edge devices, paving the way for more scalable and efficient AI solutions in various domains.