@article{3476,
  author = {Ajeet K. Jain, PVRD Prasad Rao, K. Venkatesh Sharma},
  title = {Deep Learning Optimization with MNIST and AutoEncoder Data sets},
  journal = {Journal of Information Security Research},
  year = {2022},
  volume = {13},
  number = {1},
  doi = {https://doi.org/10.6025/jisr/2022/13/1/21-28},
  url = {https://www.dline.info/jisr/fulltext/v13n1/jisrv13n1_3.pdf},
  abstract = {Optimization algorithms are extensively used in machine learning where optimization techniques are deployed. With the use of deep learning, optimization approaches are widely used with the development of new features in Stochastic Gradient Descent to convex and non-convex and derivative-free approaches. The deep learning models can able to produce systems with speed and final performance that will improve the convexity principles. Highly enhanced optimizers can increase the complexity of the depth and data sets become larger that require good optimization. Thus, in this paper, we have studied the most used optimizer algorithm in a practical way. We have experimented it in MNIST and AutoEncoder data sets. We tested in a variety of applications that can document the common features and differences and suitability of applications. Further, we have presented new variants of optimizers. Finally, we are able to find the better optimizer with the help of extensive analyses.},
}