A tech startup co-founded by a William & Mary computer scientist has been awarded a $250,000 grant from the National Science Foundation.
The NSF grant is a Small Business Innovation Research (SBIR) Phase I award to CoCoPIE, LLC. Bin Ren, assistant professor in William & Mary’s Department of Computer Science, is a co-founder of the company, which offers a novel software-based approach to endowing end-user devices with artificial intelligence (AI) functionality. For instance, the CoCoPIE website describes how the software can upscale low-resolution smartphone videos and images to high definition in real time.
Ren explained that most AI improvements are hardware-based, often coming in the form of new chips. Increases in AI functionality usually require consumers to continually buy new phones and devices to benefit. He said devices containing CoCoPIE-enriched software are not only more energy efficient, but won’t need to be replaced as often.
“CoCoPIE enables efficient AI with a software solution,” Ren explained. “CoCoPIE’s compression/compilation co-design can extend the period that a device provides satisfying performance of certain AI functionalities to users, thus reducing the frequency of purchasing new hardware. This is an immediate and effective solution to address the emerging chip crisis.”
Moreover, CoCoPIE capability can easily be installed into existing devices as part of software updates, Ren added. He said potential customers for CoCoPIE include software and hardware producers, basically anyone interested in providing advanced AI support.
Companies that may consider adding CoCoPIE capability would range from chip producers to manufacturers of end-user devices such as smartphones, robots, autonomous vehicles, and the ever-growing range of IoT (Internet of Things) things such as smart thermostats, doorbells and so on. It’s a large customer base, Ren noted.
“Mobile phone shipment alone in the past five years has been at almost 1.5 billion units a year, and IoT devices and all mobile devices combined are expected to reach more than a trillion units in the next decade,” he said. “We expect our CoCoPIE technique can run on all of these devices.”
Ren’s collaborators and partners in CoCoPIE are researchers from Northeastern University, North Carolina State University and the autonomous-vehicle firm PerceptIn. Ren says the company will use the SBIR money to ramp up CoCoPIE operations, opening new headquarters in California and hiring software engineers to advance the product. CoCoPIE was incorporated last year near Boston. Jason McDevitt, William & Mary’s director of technology transfer, says the company has a bright future.
“This technology has great potential, which is immediately apparent to anyone who takes a quick glance at the CoCoPie website,” McDevitt said. “We're pleased that CoCoPIE has received this SBIR award, and very excited about the prospects for the company.”