We are excited to announce Radical Ventures’ newest investment: Unlearn.AI.
San Francisco-based Unlearn has devised a powerful, fascinating way to apply AI to improve clinical trials. Using advanced biostatistics and machine learning, Unlearn creates “digital twins” of clinical trial participants, reducing the number of actual humans that need to be recruited for control arms. This makes clinical trials cheaper and faster to execute, ultimately enabling life-changing therapeutics to reach patients sooner.
Unlearn’s recent momentum has been incredible. The company just announced a major commercial engagement with pharmaceutical giant Merck. Several other pharma companies have also begun to deploy Unlearn’s technology. Last month the company notched a major regulatory win, with the European Medical Agency (Europe’s version of the FDA) officially voicing support for Unlearn’s digital twin technology.
Unlearn’s technology is an example of a broader trend that is becoming increasingly important in the world of machine learning: the ability to generate datasets digitally, reducing the need to laboriously collect training data from the real world. Another Radical portfolio company doing cutting-edge work along these lines is Waabi, which has developed a scalable, high-fidelity closed-loop simulator for self-driving cars named Waabi World. Waabi World is an immersive and reactive digital environment powered by AI that can design tests, assess skills, and teach the self-driving “brain” to learn to drive without ever putting a car on the road.
From Unlearn to Waabi and beyond, we believe that innovative new methods to synthetically generate valuable datasets will represent a key part of the AI technology stack in the years ahead.
Radical participated in Unlearn’s $50 million Series B alongside Insight Partners and existing investors DCVC, 8VC and Mubadala Ventures. We are thrilled to partner with the company on their journey ahead.
AI News This Week
-
Shopify makes strategic investment in US AI recommendation startup (BetaKit)
E-commerce giant Shopify announced a strategic investment in Radical Ventures portfolio company Crossing Minds. This marks Shopify’s first investment in an AI-powered recommendation platform. Merchants can now deliver individualized product recommendations without prior web history or personal data. Marketplaces are building strategies that comply with stricter consumer privacy laws, shopper concerns, and finding alternatives to third-party cookies that typically track users across the web. When asked where Crossing Minds plans to invest the capital from Shopify, CEO Alexandre Robicquet told BetaKit, “Canada, R&D, and building the team.”
-
The brain-reading devices helping paralyzed people to move, talk and touch (Nature)
Brain-computer interface (BCI) systems, most recently popularized by Neuralink, have been in experimentation for the last two decades. As early as 2006, Leigh Hochberg and his team demonstrated a patient could learn to move a cursor around a computer screen, control a television, and use robotic arms and hands just by thinking. These early implants were hardwired, cumbersome, and patients’ movements were slow and imprecise. Today’s BCI users have “much finer control and access to a wider range of skills.” Machine learning has significantly boosted the field by identifying and linking patterns to users’ intentions. Decoding neural activity is also made easier as researchers implant multiple BCIs in different brain areas. Advances in machine learning are also likely to help ensure devices function equally well for every user by enabling personalized recalibration steps.
-
Engineers enlist AI to help scale up advanced solar cell manufacturing (MIT News)
Most solar panels today are silicon-based and require high temperatures to manufacture. Perovskites are a known alternative studied globally as a lighter, thinner, and less energy-intensive alternative. However, most perovskites made today require a slow spin-coating technique that is not viable for large scale production. A combined MIT and Stanford research team is using machine learning to develop a new process that involves spraying a rolled sheet with perovskite materials. Usually, process creation requires testing many variations of inputs like humidity, temperature, processing path speed, spray nozzle distance, and curing methods. A computer model “can provide insights much faster than manual testing could ever achieve.”
-
AI and infrared spectroscopy identify the age and species of mosquitoes (Physics World)
Malaria presents a paradox: the malaria parasite, transmitted by several Anopheles mosquitoes species, requires more time to reach its human-transmissible stage than the average mosquito’s life. Therefore its persistence depends on a small proportion of mosquitoes that live long enough to transmit to a host. To assess malaria control interventions, such as insecticidal nets, it is essential to monitor the species-specific age structure of mosquito populations. Researchers combined mid-infrared spectroscopy with deep learning to develop a fast, cost-effective way to identify the species and age of three species of malaria-carrying mosquito.
-
Researchers develop a ‘bear-dar’ that warns humans of approaching polar bears (Smithsonian Magazine)
AI systems in Canada are being trained to identify polar bears. The animals are increasingly moving to the southern range of their arctic habitat due to climate change. In the northern community of Churchill, Manitoba a modified military system has been programmed with AI and trained to detect the bears from a radar system. Machine learning has helped to reduce false alarms created by humans, dogs, and other large mammals such as caribou.
Radical Reads is edited by Leah Morris (Senior Director, Velocity Program, Radical Ventures).