Mahesh Hegde
Independent Researcher
India
Abstract
Handwritten digit recognition is a fundamental problem in the field of pattern recognition and machine learning, with widespread applications in postal mail sorting, bank check processing, and automated data entry systems. Deep Neural Networks (DNNs), especially Convolutional Neural Networks (CNNs), have demonstrated remarkable success in accurately classifying handwritten digits. However, training these networks requires significant computational power and time. Graphics Processing Units (GPUs) have emerged as a powerful solution to accelerate DNN training and inference by leveraging parallelism inherent in their architectures. This study presents an in-depth analysis of GPU-accelerated deep neural networks applied to the task of handwritten digit recognition using the MNIST dataset. The work compares the training times and classification accuracies of DNNs executed on CPUs and GPUs, explores different GPU-accelerated frameworks, and discusses optimization strategies employed until 2021. Experimental results show that GPU acceleration significantly reduces training time while maintaining or improving accuracy. This paper concludes with a discussion on the implications for real-world engineering applications and future directions within the constraints of technologies available until 2021.
Keywords GPU acceleration, deep neural networks, handwritten digit recognition, convolutional neural networks, MNIST dataset, parallel computing
References
- LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324.
- Krizhevsky, A., Sutskever, I., & Hinton, G.E. (2012). ImageNet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems, 25, 1097-1105.
- Raina, R., Madhavan, A., & Ng, A.Y. (2009). Large-scale deep unsupervised learning using graphics processors. Proceedings of the 26th Annual International Conference on Machine Learning, 873-880.
- Chetlur, S., Woolley, C., Vandermersch, P., et al. (2014). cuDNN: Efficient primitives for deep learning. arXiv preprint arXiv:1410.0759.
- Deng, L. (2012). The MNIST database of handwritten digit images for machine learning research. IEEE Signal Processing Magazine, 29(6), 141-142.
- Abadi, M., Barham, P., Chen, J., et al. (2016). TensorFlow: A system for large-scale machine learning. 12th USENIX Symposium on Operating Systems Design and Implementation (OSDI 16), 265-283.
