This one-day introductory workshop dives deep. You will explore deep neural classification, LSTM time series analysis, convolutional image classification, advanced data clustering, bandit algorithms, and reinforcement learning. It’s a hands-on class; you’ll learn to implement and understand both deep neural networks as well as unsupervised techniques using TensorFlow, Keras, and Python. Just as importantly, you’ll learn exactly what types of problems are appropriate for deep learning techniques, and what types of problems are not well suited to deep learning.
The instructors for this workshop, “Deep Learning in Practice,” lead Microsoft Research’s CEO-mandated initiative to transfer deep learning intelligence into all products, services, and supporting systems across the enterprise. Workshop participants will access much of the same state-of-the-art training material used for this work at Microsoft. Along the way, the instructors will cover case studies detailing large-scale deployments for their internal clients that have generated astounding ROIs.
During the day, workshop attendees will gain the following practical hands-on experience:
- How to prepare, normalize, and encode data for deep learning systems.
- How to install deep learning libraries including TensorFlow, Keras, and CNTK, and the pros and cons of each library.
- How to create deep learning predictive systems for various kinds of data: classical business data, time series data (such as sales data), image data (such as the famous MNIST dataset for handwriting recognition), and text/document data (such as legal contracts). These datasets are a great place to start – however, for the more experienced attendee, even more challenging, “next level” datasets, such as for object recognition, will be optionally available.
