Previously, we modified the MNIST For ML Beginners Tutorial to process Mini-MIAS imagery [http://jacob.vi/modifying-tensorflows-mnist-for-beginners-tutorial-to-search-for-cancer-in-mammograms/] . The results of this were not great. Let's try this again but with a simple Multilayer Convolutional Network (CNN) in TensorFlow [https://www.tensorflow.org/versions/r0.11/tutorials/mnist/pros/…
Stanford in the news today with another deep-learning breakthrough in cancer diagnostics: Deep learning algorithm diagnoses skin cancer as well as seasoned dermatologists [http://news.stanford.edu/2017/01/25/artificial-intelligence-used-identify-skin-cancer/] First, the impressive results: > The final product, the subject of a paper in the Jan. 25 issue of…
I am presently reading Artificial Intelligence in Medicine (AAAS Selected Symposium), edited by Peter Szolovits [https://www.amazon.com/Artificial-Intelligence-Medicine-selected-symposium/dp/0891589007/ref=sr_1_1?s=books&ie=UTF8&qid=1478205510&sr=1-1&keywords=artificial+intelligence+in+medicine+peter+szolovits] . This book is an edited…
Google's TensorFlow project has posted many tutorials showing how machine learning can be used to classify images. We are going to go through these tutorials and adapt each to detecting cancer in mammography. We will start with the MNIST For ML Beginners [https://www.tensorflow.org/versions/r0.…
In order to build a machine learning model to help us with mammography-based diagnosis[1] we will need data to train our models. Thankfully, there are a number of data sets publicly available. University of South Florida Digital Database for Screening Mammography [http://marathon.csee.usf.edu/Mammography/Database.html]…
According to Mattermark, No. AI Is The New New Thing (And It's Taking Over) [https://mattermark.com/ai-new-new-thing-taking/] This editorial presents an interesting look at the ecosystem of machine learning startups, across all industries. Some choice quotes: > 2013’s Big Data is 2016’s Machine Learning (ML)…
Microsoft is taking their Machine Learning prowess and applying it to research into cancer care. How Microsoft computer scientists and researchers are working to ‘solve‘ cancer [http://news.microsoft.com/stories/computingcancer/] While some of their approaches are at the basic science level, many are applicable to a practicing oncologist.…
Machine Learning (ML) is a technique to train computers to identify patterns instead of explicitly telling them what those patterns are. If you want to teach a machine-learning system to recognize a cat, you don't tell it to look for whiskers, ears, fur, and eyes. Instead, you show…