Checking in with the Intel® AI Lab

The neon™ deep learning framework was created by Nervana Systems to deliver industry-leading performance. As of 2018, the neon framework is no longer being supported. We recommend customers to consider Intel optimized frameworks listed here.

Intel’s Artificial Intelligence Products Group has had a busy year. Last month, we announced the year-end  availability of the Intel Nervana™ Neural Network Processor, the first in a family of processors designed from the ground up for AI workloads. A few days later, we released the Reinforcement Learning Coach, an open source research framework for training and evaluating reinforcement learning (RL) agents. And just a few weeks ago, we released the latest generation of neon™, our open source deep learning framework. On Monday, December 4th at the NIPS conference, we presented our paper on Flexpoint, an innovative numerical format for the efficient training of deep neural networks. In addition, we will be presenting our work on “Unsupervised Deep Structure Learning by Recursive Independence Testing” at the Bayesian Deep Learning workshop on Saturday, December 9th. We will also present our 3rd place winning work in the Defense Against Adversarial Attack competition and our 4th place winning work in the Targeted Adversarial Attack competition. These exciting developments are all made possible by the Intel® AI Lab.

Intel AI Lab, a team of machine learning and deep learning researchers and engineers, data scientists, and neuroscientists, has been focused on state-of-the art research and development in the field of artificial intelligence. The lab was formed earlier this year, bringing together AI teams from Nervana Systems and from across Intel. Our core areas of research range from projects with a direct impact on upcoming and future generations of Intel technologies to novel algorithm development in areas such as  natural language, speech, and vision. Together with our full stack and optimization expertise, we often apply the results of our research to help our customers solve their business problems. Through our unique team composition we are also exploring integrative approaches to AI spanning across traditional domain and task boundaries.

The lab collaborates with academic research institutions and corporations to solve problems using AI. These collaborations include researchers at the University of California Berkeley, Brown University, CERN, and Bar Ilan University, and companies from industries such as finance, energy, healthcare, retail, manufacturing, automotive, government, and media & entertainment. We are committed to strong ethical principles and the use of AI for good, and are deeply involved with the Partnership on AI to benefit people and society.

Advancements in AI Research

By developing state-of-the-art algorithms, our team is building models and applications in areas including natural language processing, computer vision, autonomous driving, speech recognition, personalization, anomaly detection, and robotic learning. Many of the standard and benchmark models are developed and provided using our open source neon™ or Intel® nGraph (formerly known as Intel® Nervana Graph) AI software platforms and are available on the Model Zoo or as examples in neon. These frameworks take deep learning models through compilation and optimization to run efficiently on a variety of hardware platforms. The models we provide demonstrate the usage and flexibility of Intel frameworks, and enable deep learning researchers and developers to leverage the most relevant topologies and Intel technologies for their use cases. As we move forward with more original research, we will continue to publish papers and open source models. In the long run, we anticipate that the next phase of AI software will require the creation and proliferation of reusable, modular capabilities to bridge the monolithic topology and framework abstractions of today.

Out of the Lab, Into the Real World

We’ve leveraged our AI and data science research for enterprise use cases, including knowledge management for the financial services industry, in-car speech recognition, smart farm machines for agriculture, advanced analytics for motorsports, product recommendations for retail, underwater infrastructure inspection for an energy company, trading strategy optimization for asset management, and video analysis and NLP for media. Today, we are excited to announce our collaboration with Synthetic Genomics (SGI), an industry leader in genomics research and bioproduction. We worked with SGI to develop novel architectures to parse genetic code and help solve puzzles like predicting protein properties from amino acid sequences. To do this, we designed a topology that leveraged deep learning capabilities for analysis, classification and prediction.

We have also applied deep learning to our own manufacturing processes to better predict faults at an early stage of production. We demonstrated that deep learning techniques provide a viable alternative to statistical analysis platforms for manufacturing fault detection applications, and can benefit from being combined with traditional models to achieve the highest possible accuracy. In these applications, each percentage of accuracy gained has significant financial impact for Intel.

The Intel AI Lab team is actively participating in and driving leading edge AI research, powered by Intel software and hardware technologies. We believe that AI will have a transformative impact on our society, and we’re proud to be part of Intel’s AI portfolio that will advance adoption and innovation. If you’re attending the NIPS conference this week, stop by our booth to learn more about the exciting activities we’re working on – and how you can join the team. If you’re not at NIPS, you can also reach us at