Paper Title
Identification of Cancer Subtypes with Deep Learning in Breast Cancer

Abstract
Cancer is the most dangerous disease that can affect people around the globe. World health organization has reported 68500 and 2.3 million with breast cancer globally. Some difficulties are identified in cancer while detecting histopathology images. However, manual detection is a very costly and tedious job. The major causes of breast cancer are excessive body weight, physical inactivity, and consumption of alcohol. Many of the recent approaches follow Machine Learning algorithms that are based on statistical data for identifying cancer. One of the major drawbacks of a machine learning algorithm it is difficult to identify data patterns in large amounts of data sets as it contains noisy data. The traditional methodologies for cancer detection lack imbalanced data, over-fitting, and scalability-related problems. To overcome this problem, On research, it shows deep learning neural networks are sensitive to adversarial networks. Furthermore, some of the manipulated samples can show incorrect predictions by using deep learning neural networks. This model is again subdivided into two parts namely variational auto-encoder(VAN) and general adversarial neural networks(GAN). These techniques are used to detect cancer at four stages. Based on statistical data representations, we group patients into various cancer sub-types. Keywords - Deep Learning Neural Networks, Histopathology Images, Variational Auto-Encoder, And Generative Adversarial neural networks.