Transfer Learning on Deep Neural Network: A Case Study on Food-101 Food Classifier

—In the era of the ‘Big-Data’ we hear a lot about machine learning for working on this big data. Machine learning helps us to predict and analyze data with better accuracy and least human intervention. Machine learning is autonomous but susceptible to errors. This is due to biased prediction when previously trained on small data. This leads to chain of errors that can`t be determined easily for long period of time. And when recognized takes lot time to recognize source. There comes the idea of deep learning which achieves the flexibility by using use nested hierarchy of concept to define the world. But deep learning has setback of taking very long time to train data which could be reduced by using transfer learning.