Random Resistive Memory-Based Deep Extreme Point Learning Machine

The rapid advancement of artificial intelligence has driven the development of more efficient and scalable machine learning models. In this paper, we propose a novel approach combining random resistive memory with a deep extreme point learning machine for unified visual processing. Resistive memory, such as memristors, offers non-volatile, low-power memory storage that is well-suited for neuromorphic computing, enabling enhanced energy efficiency and parallel processing capabilities. By integrating this memory into a deep learning framework, we introduce a new model architecture that reduces training time and computational overhead. The extreme point learning mechanism optimizes learning by focusing on critical data points, improving the model’s ability to generalize across a wide range of visual tasks. This unified model processes multiple visual tasks simultaneously, including image classification, segmentation, and object detection, with a high level of accuracy and efficiency. Experimental results demonstrate the effectiveness of our proposed architecture in terms of both performance and energy consumption, highlighting its potential for real-time, large-scale visual processing applications. This work lays the groundwork for future research into hardware-efficient deep learning models and their applications in computer vision.