DARWIN-Intel Combination Delivers 16.3X Speedup on Image Classification Networks, Additional Inference Performance Improvements
Product News Jul 17, 2019
Image credit: DarwinAI
DarwinAI, a Waterloo, Canada startup creating next-generation technologies for Artificial Intelligence development, announced that the company’s Generative Synthesis platform – when used with Intel technology and optimizations – generated neural networks with a 16.3X improvement in image classification inference performance. Intel shared the optimization results in a recently published solution brief.
“The complexity of deep neural networks makes them a challenge to build, run and use, especially in edge-based scenarios such as autonomous vehicles and mobile devices where power and computational resources are limited,” said Sheldon Fernandez, CEO of DarwinAI. “Our Generative Synthesis platform is a key technology in enabling AI at the edge – a fact bolstered and validated by Intel’s solution brief.”
DarwinAI and Intel Optimizations Deliver As Much As 16.3X Performance Increase
Intel engineers ran image classification performance tests with ResNet50 and NASNet, two popular neural networks, using Intel Optimizations for TensorFlow with Intel Math Kernel Library (Intel MKL and Intel MKL-DNN). The results: for ResetNet50, Darwin’s Generative Synthesis platform combined with Intel’s optimizations delivered a 16.3X improvement in inference speed over baseline measurements on an Intel Xeon Platinum 8153 processor. Meanwhile, for NASNet, the DarwinAI-Intel combination yielded a 9.6X improvement in inference speed on the same hardware.
“AI Building AI” Technology Produces High-Performance Deep Learning Solutions
DarwinAI’s Generative Synthesis platform uses AI itself to examine and learn from a neural network in order to construct new highly compact versions without sacrificing functional accuracy. This patented “AI building AI” technology dramatically reduces the size, complexity and guesswork in designing efficient, high-performance deep learning solutions. This unique method of understanding also facilitates “explainable” deep learning -- the ability to understand why a network makes the decisions it does.
The benefits of DarwinAI’s Generative Synthesis technology include:
- Faster, smaller models for deploying AI at the edge
- Accelerated design through human-machine collaboration
- AI-powered profiling tools that identify performance bottlenecks
- “Root cause analysis” features via the platform’s AI-powered explainability tools -- specifically, the ability to identify errors and biases in the data, and key factors most influencing a particular decision
DarwinAI is a member of the Intel AI Builders Program, an ecosystem of industry-leading independent software vendors (ISVs), system integrators (SIs), original equipment manufacturers (OEMs), and enterprise end users, which have a shared mission to accelerate the adoption of artificial intelligence across Intel platforms.
Continued Company Momentum, Industry Awards and Accolades
DarwinAI’s impressive optimization results with Intel are the latest in a series of industry honors and accolades. The company, its team and award-winning technology have been recognized by leading publications, research organizations and analyst firms:
- NeurIPS (Neural Information Processing Systems) conference – presented five papers at the 32nd NeurIPS conference in 2018, the most prestigious academic conference for deep learning
- Hello Tomorrow 2019 – named a “Top 500 Deep Tech Startup” (from more than 4,500 applications) and named one of seven finalists in the "Data and AI" track
- InsideBIGDATA – named to the publication’s IMPACT 50 list of “movers and shakers” twice, Q1 2019 (#48) and Q2 2019 (#44)
- Frost and Sullivan – earned a 2019 Technology Innovation Award for its research and development efforts
- Timmy Awards – recognized as one of Ontario's top tech startups at the 4th annual
- Timmy Awards (runner up in "Best Tech Startup" category
- DesignNews – named one of the “12 Automotive Startups to Watch in 2019”