Deep Residual Network using Exponential Linear Unit
Designed a model for image classification which improves the learning behavior of Residual Networks and has better accuracy compared to previously published state-of-the-art models with 5.62% error rate, least in existing models The preprint arXiv paper has been cited 10 times and received appreciation from Node.js founder, Ryan Dahl via Twitter Publication: ACM Digital Library ISBN 978-1-4503-4301-5
