Educational outreach inspires and motivates a new generation of AI practitioners schooled in the latest technologies and making the most of cost-effective solutions.
“Intel is democratizing AI by providing a low-cost, easy-to-use platform with the Intel® Movidius™ Neural Compute Stick.”
— Oliver Chen, maker and educator, Intel
As artificial intelligence (AI) and Internet of Things technologies continue to advance and make deeper impacts throughout the world, new skills and knowledge are needed, provided by expanded educational opportunities, to help the next generation move forward and explore the possibilities.
A curriculum designed to capture the imaginations of young people and equip them with programming and design skills to use AI in innovative ways offers a path to rewarding work and lifelong careers. The AI Digital Literacy project, developed by Oliver Chen at the Sacramento Public Library, is an exciting step forward on this path.
Oliver Chen is a technical marketing engineer at Intel who provides customer support for AI products. He is a strong proponent of educating students in science, technology, engineering, and math (STEM) and contributes to numerous programs and projects to get young people energized and motivated toward working in these fields. His career reflects a diverse range of experience, including engineering, enterprise information technology, product development, and general management.
A recent engagement with the Sacramento Public Library (SPL), funded by a Winners of Wonder grant sponsored by Intel to celebrate the company’s 50th anniversary, led to the development of an AI curriculum for teaching coding skills to first-timers. Oliver extended Scratch*, a visual programming tool for beginners, to operate with the Intel® Movidius™ Neural Compute Stick (Intel® Movidius™ NCS). Scratch was developed by the Lifelong Kindergarten Group at the MIT Media Lab and is free for anyone to use.
“Using a set of about 100 commands that can be snapped together visually, you can create just about anything and learn the fundamentals of more advanced languages. In addition, there’s a whole community of people online who will give you feedback on your projects and gladly help you with your questions.”1
— 14-year-old Scratch user
SPL is using the curriculum that Oliver developed to help provide students in the South Natomas area of Sacramento with essential coding skills. This project is part of an SPL initiative—known as the South Natomas Innovation Station—designed to boost digital literacy in the area and provide a wide range of useful skills to the community. The curriculum is also being used by the NorCal Girl Scouts and the Boy Scouts Order of the Arrow national honor society, two groups that are based in rural regions.
Figure 1. The Sacramento Public Library collaborated with Intel on the development of the AI curriculum.
“My Winners of Wonder project,” Oliver said, “helps enable digital literacy through an AI curriculum that is easy for first-time programmers to consume, easy for educational outreach venues to maintain, and at a low price point to enable broader adoption. The project combines using the Raspberry Pi* and the Intel Movidius NCS through Scratch and Python*. The content will be used by the Sacramento Public library to reach the general public, and by the Girl Scouts Heart of Central California (GSHCC) to focus on bringing STEAM (science, technology, engineering, arts, and math to girls).”
Oliver spent time as a software mentor, working with young innovators, and was impressed by how the inexpensive, accessible Arduino*-compatible boards inspired numerous, ingenious student projects. “I now support Intel’s AI products to external customers,” he said, “and I hope to help build a similar great experience for those wanting to learn about AI.” “I have been working on this project,” Oliver noted, “ever since the grant was awarded in August during Intel’s 50th Anniversary Celebration, and I should have it completed by the end of November 2018.”
“Fueled by the desire to help struggling stroke patients regain their quality of life, I invented MotivateMe*, a smart wristband that gives stroke patients real-time, motivational feedback on their rehabilitation exercise performance. I presented my research to President Obama at the 2016 White House Science Fair, at the Intel Developer Forum (IDF) conference, and at the World/ National/Bay Area Maker Faires*. Teaching myself computer science and electrical engineering skills empowered me to solve a real-world problem, and I founded STEAM Ahead to encourage and equip other students to tackle the social issues close to their hearts.”2
— Diana Voronin, founder of STEAM Ahead
Figure 2. Intel Shooting Star drones aligned in celebration of the company’s 50th anniversary.
“Don’t be encumbered by history. Go off and do something wonderful.” - Robert Noyce, Intel founder
As he began this project, Oliver relied extensively on the Intel® Movidius™ Software Development Kit (SDK), available on GitHub*. The custom Scratch blocks that were created to give students an easy way to begin creating applications with the Intel® NCS were based on the Neural Compute Application Zoo (NC App Zoo).
Oliver recommends that innovators interested in taking the next step to produce an actual product—beyond just experimenting and prototyping with the Intel Movidius NCS—port their applications easily to should consider the Intel® Movidius™ Vision Processing Unit (VPU)-based solutions from one of Intel’s AI: In Production Program partners.
He also gave this advice for developers getting started with AI projects: “Intel® AI DevCloud offers free cloud-compute resources to Intel® AI Academy members for their machine learning needs. Also, developers looking for funding should consider government and corporate grants for research in AI. Beyond financial assistance, some programs provide cloud- compute credits and data labeling services.”
“There’s certainly a significant under-representation problem in all areas of STEM, and this is higher in computer science than in some other areas and higher in artificial intelligence than in general computer science. So, diversity is definitely a significant problem. And there’s a tremendous shortage of skilled people to do this work. There’s this general cry of ‘We need more people, so we really can’t afford to leave talent on the table.”3
— Jennifer Rexford, Computer Science Chair, Princeton
“I am encouraged when I see companies like Intel and the Raspberry Pi Foundation collaboratively work on projects to inspire younger AI innovators,” said Oliver. The Raspberry Pi Foundation hosts a number of learning resources for new programmers on the FutureLearn site.
Figure 3. Early engagement with technology can be fun even for very young students.
A key component of Oliver’s project is the Intel Movidius Neural Compute stick. This small, low-power AI inference accelerator in a USB form factor can perform a wide variety of AI application tasks locally, without requiring a cloud connection. As such, it complements the AI curriculum that Oliver developed and that SPL is implementing in Sacramento. Through the extensions that were added, students using Scratch can take their education a step further and begin exploring the real-world possibilities offered by deep learning applications in a familiar, beginner’s programming environment. Custom Scratch blocks are provided, based on the NC App Zoo, to streamline the learning process.
Figure 4. The Intel® Movidius™ NCS supports offline neural network applications.
Once young developers gain a basic understanding of coding principles and AI applications, the Intel® Movidius™ SDK provides a platform for building convolutional neural networks using the Intel Movidius NCS, with support for Python* and a framework for image classification. The Intel Movidius NCS is particularly well suited for vision applications where a continuous cloud connection is impractical or impossible.
Two deep learning frameworks are currently supported by this platform: TensorFlow* and Caffe*. The SDK can effectively convert models created in these frameworks to a format that can run on the Intel Movidius NCS. Three libraries contained in the SDK make this possible. In a video explaining the capabilities of the Intel Movidius Neural Compute Stick, Siraj Raval, director of the School of AI in San Francisco, said, “Compile converts a model into the appropriate format. Profile gives layer-by-layer statistics to evaluate model performance. Check compares the inference results from running the network on the device versus pure TensorFlow or Caffe” (see Figure 5). This same video provides information about how to take advantage of the community projects in the NC App Zoo on GitHub, which can be cloned as a starting point for new projects involving the Intel Movidius NCS.
Figure 5. Components of the Intel Movidius SDK help convert models from other frameworks.
The Intel® Distribution of OpenVINO™ toolkit also supports the Intel Movidius NCS with a number of pre-trained models that support diverse vision applications, including face and emotion detection, vehicle detection, pedestrian and bike detection, and innovative uses in smart cameras and robotics. Software developers and data scientists who want to accelerate solutions across a wide range of platforms can take advantage of the Intel Movidius NCS support for CPUs, GPUs, VPUs, and FPGAs, using the same code for each platform.
AI applications running on the Intel Movidius NCS at the edge of the network have the benefit of additional security, since data and operations are all performed locally, reducing the risk of exposure. The Intel Movidius NCS also reduces latency issues by eliminating the need to remain in close communication with the cloud.
The Intel Movidius NCS also reduces latency issues by eliminating the need to remain in close communication with the cloud. The Intel Movidius NCS provides a gateway into the world of AI development, giving students a clear understanding of how applications can be built to run in a real-world environment. The hope is that this technology will spark the imagination of young students—once they see the possibilities—and encourage them to pursue innovative projects to help people and businesses enhance everyday life.
Intel recently introduced the next generation AI development kit—the Intel® Neural Compute Stick 2 (Intel® NCS 2). Based on the latest Intel Movidius Myriad X VPU, the Intel NCS 2 delivers AI inference acceleration in the same USB form factor but with increased SHAVE cores and a dedicated Neural Compute Engine boosts performance over the previous generation.
Through the design and development of specialized chips, research, educational outreach, and industry partnerships, Intel is accelerating the progress of AI to solve difficult challenges in medicine, manufacturing, agriculture, scientific research, robotics, and other industries. Intel works closely with policymakers, educational institutions, and enterprises of all kinds to uncover and advance solutions that address major challenges in the sciences.
Figure 6. Educational venues in many communities are beginning to include AI in their curricula. The Intel® AI Portfolio includes: Reinforcement Learning Coach: Provides an open source research framework for training and evaluating RL agents by harnessing the power of multicore CPU processing to achieve state-of-the-art results.
The Intel® AI Portfolio includes:
Reinforcement Learning Coach: Provides an open source research framework for training and evaluating RL agents by harnessing the power of multicore CPU processing to achieve state-of-the-art results.
Framework Optimization: Achieve faster training of deep neural networks on a robust scalable infrastructure.
Intel® Xeon® Scalable processors: Tackle AI challenges with a compute architecture optimized for a broad range of AI workloads, including deep learning.
Intel® Movidius™ Myriad™ X Vision Processing Unit (VPU): Delivers advanced features for the most demanding computer vision workloads and deep neural network implementations.
Intel® Movidius™ Neural Compute Stick: Provides deep learning prototyping at the network edge with always-on vision processing making it ideal for use in smart security cameras, gesture controlled drones, industrial machine vision equipment, and more.
Intel® (FPGA): Create specialized, custom functionality for a wide variety of electronic equipment, including AI-based solutions and monitoring devices, medical equipment, aircraft navigation devices, system accelerators, and more.
Intel® Distribution of OpenVINO™ toolkit: Make your vision a reality on Intel® platforms—from smart cameras and video surveillance to robotics, transportation, and more.
Intel® Distribution for Python*: Supercharge applications and speed up core computational packages with this performance-oriented distribution.
Intel® Data Analytics Acceleration Library (Intel® DAAL): Boost machine learning and data analytics performance with this easy-to-use library.
Intel® Math Kernel Library (Intel® MKL): Accelerate math processing routines, increase application performance, and reduce development time.
For more information, visit the portfolio page.
Intel's compilers may or may not optimize to the same degree for non-Intel microprocessors for optimizations that are not unique to Intel microprocessors. These optimizations include SSE2, SSE3, and SSSE3 instruction sets and other optimizations. Intel does not guarantee the availability, functionality, or effectiveness of any optimization on microprocessors not manufactured by Intel. Microprocessor-dependent optimizations in this product are intended for use with Intel microprocessors. Certain optimizations not specific to Intel microarchitecture are reserverd for Intel microprocessors. Please refer to the applicable product User and Reference Guides for more information regarding the specific instruction sets covered by this notice.
Notice revision #20110804