Radical Reads — September 28, 2020

Original article was published by Radical Ventures on Artificial Intelligence on Medium


Radical Reads — September 28, 2020

Curated deep tech and AI content that humans at Radical are reading and thinking about. Sign up here to have Radical Reads delivered directly to your inbox every week.

1) AI in Fintech: Sensibill launches smart technology for rapid receipt data extraction (BusinessWire)

“Sensibill, the leading provider of SKU-level data and financial tools like digital receipt management that help institutions better know and serve their customers, today announced the launch of its newest product: Receipt Extraction API. The machine learning-based solution rapidly automates and streamlines the transcription of receipts, allowing businesses to deepen customer engagement and loyalty at scale.”

Radical Commentary: Sensibill, a Radical portfolio company, launched a machine learning solution to automate receipt transcription. Sensibill’s Receipt Extraction API solution stands to benefit a wide range of businesses that need to quickly and accurately extract receipt data at scale like enterprise accounting firms, financial services companies, and loyalty and reward companies.

Digital transformation is a high priority across the financial services sector, which now have fewer face-to-face interactions with their customers. Digitization of financial data permits these organizations to reduce traditional manual processes in their workflow. Adding machine learning enables a deeper level of understanding of customers. Despite the challenges presented by the pandemic, global digital transformation expenditure in the financial sector is forecast to grow at 13.5% this year.

2) Strengthening Canada’s AI Ecosystem: University of Toronto receives single largest gift in Canadian history from James and Louise Temerty to support advances in human health and health care (University of Toronto)

“The transformational gift from the Temerty Foundation, established by James and Louise Temerty, will support advances in machine learning in medicine; biomedical research and collaboration across Toronto’s health-science network; innovation, commercialization and entrepreneurship; equity and accessibility in medical education; and the creation of a new state-of-the-art Faculty of Medicine building for education and research.”

Radical Commentary: In our Radical Talks podcast with Dr. Eric Topol we discuss how Canada is positioned to become a leader in AI-driven “Deep Medicine” as it boasts a high concentration of both AI researchers and healthcare researchers, an incredibly diverse population and a data-rich single-payer health system. The University of Toronto has a long history of healthcare innovation, including the discovery of insulin and the development of the world’s first artificial pacemaker.

This donation to the University of Toronto’s medical faculty includes a focus on AI, with the establishment of a new Centre for AI Research and Education in Medicine. The Temerty Faculty of Medicine is located directly across the street from 5 world-renowned research hospitals, one of the world’s largest tech hubs, the 1.5 million square foot MaRS (which houses the Vector Institute for AI, one of the world’s largest and leading AI research institutes, started by founders of Radical) and the forthcoming Schwartz Reisman Innovation Centre that will have a strong AI and healthcare focus.

There are already a significant number of cross-appointments of faculty and researchers at these institutions and hospitals. We expect this donation will bolster Canada’s position as a global leader in AI healthcare innovation by advancing the science and increasing the number of trained AI healthcare researchers, both of which will draw founders, companies and investors to enhance an already burgeoning healthcare startup ecosystem. Radical’s current fund has already made 5 investments in healthcare and we expect to continue focusing on this space which is being transformed by AI.

3) AI in Healthcare: Artificial Intelligence Detects Arthritis Before it Develops(UPMC)

“When doctors look at these images of the cartilage, there isn’t a pattern that jumps out to the naked eye, but that doesn’t mean there’s not a pattern there. It just means you can’t see it using conventional tools.”

Radical Commentary: The Centre for Disease Control and Prevention estimates that 1 in 4 adults in the US will develop some form of arthritis. Osteoarthritis is the most common form of arthritis, affecting over 32.5 million adults in the US. Despite its prevalence, doctors have long struggled to identify at-risk patients before symptoms present themselves.

This week, researchers at the University of Pittsburgh and Carnegie Mellon announced the creation of a machine-learning algorithm that can detect subtle signs of osteoarthritis on an MRI scan taken years before symptoms begin. The algorithm predicted osteoarthritis with 78% accuracy from MRIs performed three years before symptom onset. The model was trained using knee MRIs of thousands of people over seven years to see how osteoarthritis of the knee develops.

AI is proving to be an essential tool in the future of preventative medicine. It may soon be commonplace to provide at-risk patients with preventative plans including medications in an effort to avoid joint-replacement surgery, the most common surgery in the US for people over the age of 45. In the nearterm, these AI-driven insights may also prevent patients from developing rheumatoid arthritis, a related condition that can be treated.

4) AI in Drug Discovery: Artificial intelligence in COVID-19 drug repurposing(The Lancet)

“AI has been revolutionising drug discovery by extracting hidden patterns and evidence from biomedical data. …The pandemic is a good opportunity for introducing advanced AI algorithms combined with network medicine for drug repurposing. …To date, AI’s potential ability to identify new candidate therapies that can be made available for clinical trials rapidly and, if approved, merged into health care is unparalleled, making AI a centrepiece of advanced technologies. Because of this, AI is a promising method for accelerating drug repurposing for human diseases, especially emerging diseases, such as COVID-19. With the availability of big data, including biological, clinical, and open data (scientific publications and data repositories) [sic], novel AI techniques capable of leveraging these large sets of biomedical data are in high demand.”

Radical Commentary: Radical’s healthcare thesis touched on how drug discovery has been driven by an increased trust in AI solutions, as well as more sophisticated digital infrastructure. Given the high attrition rates, substantial costs, and low pace of de-novo drug discovery, exploiting known drugs can help improve their efficacy while minimising side-effects in clinical trials.

AI is a promising method for accelerating drug repurposing especially for emerging diseases like COVD-19. We expect future successful AI models, coupled with big data, to substantially improve drug repurposing and aid medical decision making of therapeutic benefits with real-world evidence for complex human diseases. As the authors of this paper put it, successful AI models in the future will be “accurate in terms of the generated outcomes, integrative of disparate information types and sources, interoperable in diverse deployment settings, interpretable of internal working mechanisms, and robust to noise and adversarial attacks.”

5) AI in the Real World: Watch a Robot AI Beat World-Class Curling Competitors (Scientific American)

“Deep-learning techniques that are all the rage in AI log superlative performances in mastering cerebral games, including chess and Go, both of which can be played on a computer. But translating simulations to the physical world remains a bigger challenge.

A robot named Curly that uses “deep reinforcement learning” — making improvements as it corrects its own errors — came out on top in three of four games against top-ranked human opponents from South Korean teams…

One crucial finding was that the AI system demonstrated its ability to adapt to changing ice conditions. “These results indicate that the gap between physics-based simulators and the real world can be narrowed.””

Radical Commentary: Deep reinforcement learning (DRL) is helping robotic systems bridge the gap between simulations and interactions in the real world. Radical’s portfolio company Covariant AI does this for pick-and-place applications in distribution centers. “Curly” is an application of similar principles to the sport of Curling.

While we do not expect a competitive robot curling league anytime soon, the robot’s ability to adapt to changing ice conditions in real time shows how DRL will be widely applicable and enable hundreds of use cases where robots interact with changing environments by learning from trial and error.

***
Editor’s Note: We will continue to use this platform to share without commentary articles focused on data and the use of it to illustrate and illuminate racial injustice. Because you cannot fix problems you cannot see or understand.

6) A Neighborhood’s Race Affects Home Values More Now Than in 1980(Bloomberg CityLab)

“U.S. fair housing laws passed in the 1960s and ’70s were supposed to help bring racial parity to a housing market that since its beginning confined Black homebuyers to the cheapest forms of housing in the most undesirable neighborhoods. But since those laws were passed, the disparity in the appraised values between homes in majority-white and predominantly non-white neighborhoods has widened dramatically, according to a new study

The new study finds that the racial composition of a neighborhood was an even “stronger determinant” of a home’s appraised value in 2015 than it was in 1980, to Black homeowners’ increasing disadvantage. Analyzing reported home values, Howell and Korver-Glenn found that the race appraisal gap has doubled since 1980: The difference in average home appraisals between neighborhoods that are majority-white and those that are predominantly Black and Latina was $164,000 in 2015, up from about $86,000 in 1980.”