09:30-10:00: Intel Machine Learning Tools for Software Developers and Data Scientists Now you can scale your machine learning and deep learning applications quickly – and gain insights more efficiently – with your existing hardware infrastructure. Popular open frameworks newly optimized for Intel, together with our advanced math libraries, make Intel® Architecture-based platforms a smart choice for these projects. This session will give an overview of software development tools from Intel for Machine Learning including Deep Learning.
10:00-11:00: Intel Performance Libraries for Machine Learning and Deep Learning Gennady Fedorov, Intel Deep learning and machine learning are nowadays used in many applications that require intensive computation. Intel software technologies can accelerate the computation intensive solutions in these applications with Intel performance libraries. This presentation will focus on two Intel performance libraries: Intel® Math Kernel Library and Intel® Data Analytics Acceleration Library, which offer optimized building blocks for data analytics and machine learning algorithms. This session will introduce the set of new features in these libraries that are specially tuned for deep learning, then will cover the newly released open-sourced MKL-DNN library that provides primitives optimized on Intel architectures for most popular deep learning frameworks and finally we will show how to get the capability and performance advantages of these libraries with the handwritten digit recognition application.
11:00-12:00: Data Analytics with Intel® Distribution for Python and pyDAAL Frank Schlimbach, Intel Intel® Distribution for Python*, powered by Anaconda, gives you ready access to tools and techniques for high performance to supercharge your Python applications on modern Intel platforms. The Intel Distribution delivers an easy-to-install, performance-optimized Python experience to meet your most demanding requirements. Among other packages for data analytics and machine learning Intel® Distribution for Python provides pyDAAL, the Python API of Intel Data Analytics Acceleration Library. The tutorial will provide a quick introduction to pyDAAL features and the API for Python for developers who are already familiar with basic concepts and techniques in machine learning.
12:00-13:00: Lunch Break
13:00-13:30: Intel optimized Machine Learning Frameworks Intel is committed to optimize popular ml/dl frameworks for Intel Architecture. We have forked off a version of Caffe, optimized with MKL/MKL-DNN for Intel Architecture and made it available through GitHub. It is significantly more performant than the mainstream. The GitHub site includes all the source and libraries you need to get started. We have also optimized other frameworks for Xeon (AVX2 and greater – HSW/BDW +) and Xeon Phi™ Processors (AVX-512 and greater – KNL +). Additional roadmaps for neon, MXNet, Spark and CNTK will be provided soon.
13:30-14:30: Distributed Machine Learning – Image Classification for Supercomputers Michael Steyer, Intel Parallelism is nothing new in the field of Machine Learning and especially Deep Learning using Convolutional Neural Networks for image classification. This presentation will focus on image classification using the Caffe framework across distributed machines. The Intel Caffe implementation, leveraging the Deep Learning capabilities of the Intel MKL library (MKL DNN) will be used to demonstrate image classification. The DNN training will be demonstrated across several different compute nodes (MPI).
14:30-16:00: Nervana demo
16:00-16:30: Coffee Break
16:30-17:30: Saffron Reasoning Systems (TBD) Reasoning Systems – TBD