Sanjana Basu, Investor

    .   

Photo Source: Radical Ventures
Notes: T
he map covers a combination of AI first and applied AI companies. Companies may belong in more than one category or in different categories. 

Digital healthcare adoption has accelerated since the onset of the pandemic. In the first half of 2021, approximately $14.7B was invested in digital healthcare in the US, surpassing the total amount invested in 2020 (~$14B). Building on a world-class talent pool, AI startups are also booming in Canada with ~$1.4B raised in the first 6 months of 2021, up 190% YoY.

Radical’s Canadian AI and Healthcare startup market map aims to capture this momentum.

The companies in the map fall into 3 main categories:

(1) AI enabled decision support: Companies in this category are mapped across the value chain of care including: diagnosis and screening, treatment selection and planning, and remote monitoring. Companies could belong to more than one part of this value chain and have been categorized for the purpose of clarity.

Diagnosis and screening is the most active space for AI companies. Most companies use imaging as a modality for diagnosis while others use clinical data captured in different formats to detect biomarkers. Treatment selection and planning, as well as remote monitoring, augment both the clinician and the patient. The most active indications include mental health, gut health, fertility, neurodegenerative diseases, and chronic conditions like diabetes and cardiac health. Emerging demographic trends include women and children’s health. Finally, prescription digital therapeutics also fall under this category and can be considered a companion or alternative to traditional therapeutics.

(2) AI accelerating therapeutics: The acute urgency for developing treatment solutions during the pandemic sets the stage for inventions that fast track the medicine to market process.

Over the last few years, AI for drug discovery and development companies have started seeing AI-discovered drug candidates through to clinical development to capture the value of the assets. Another set of companies continue to focus on research intelligence by leveraging software to enable discovery and medical research. Finally, the prediction power of AI is used extensively today in clinical trial technology to optimize recruitment, saving millions of dollars and significant time costs for pharmaceutical companies.

(3) Data infrastructure and process efficiencies: Companies in this category sell primarily to providers to build process efficiencies across clinical workflow and revenue cycle management tools. These companies are also building data infrastructure to ensure data security, enable data interoperability, and set the foundation for a new generation of technology-first healthcare products.

Canada has a strong foundation to develop new healthcare technologies given the nexus of world-class health and machine learning researchers and a single-payer data source. Radical continues to play an active role in the ecosystem, commercializing, supporting, and scaling transformative innovations in healthcare.

If we have missed your company, please reach out to sanjana@radical.vc.

5 Noteworthy AI and Deep Tech Articles: week of August 9, 2021

1) An endlessly changing playground teaches AIs how to multitask (MIT Technology Review)
AIs benefit from playgrounds just like children. Instead of developing the skills needed to solve a particular task, ‘playgrounds’ (real or virtual) enable experimentation and exploration. Open-ended learning can allow AIs to generalize skills to succeed in tasks that they have not previously seen, which is an important step toward generalized intelligence. 

2) Deep learning for pulmonary embolism detection (Nature)
Pulmonary embolism (PE) is associated with significant morbidity and mortality. Prompt and accurate diagnosis can substantially reduce mortality, and deep learning models may be able to help. The current diagnostic method is time consuming and requires radiologists’ expertise. Future research on the strengths and limitations of these algorithms is needed as deep learning models are gradually being implemented in hospital systems.

3) How AI Will Help Keep Time at the Tokyo Olympics (Wired)
Timing Olympic races is a niche space for innovation. Omega has been innovating on this task for more than 85 years. Time-keeping has come a long way from hand-stopping and now uses cameras with computer vision technologies to track not only athletes, but also the ball in some sports. In volleyball, ensuring the computer accurately tracks the ball when the cameras can no longer see it has been a major hurdle which predictive AI helps solve in real-time. Omega is looking to introduce additional AI-powered technologies in time for the 2024 Paris Games.

4) Fake It to Make It: Companies Beef Up AI Models With Synthetic Data (Wall Street Journal – subscription required)
Detecting fraud depends on an ability to flag suspicious behaviour patterns. When fraud is committed in an unexpected or new way, detection can be difficult. Machine learning and data scientists have been experimenting with synthetic data in hopes of improving AI-based fraud-detection models.

5) Artificial intelligence vs. neurophysiology: Why the difference matters (VentureBeat)
Turns out AI predictions and human reflexes share one big thing in common: they are both significant energy consumers. The algorithms that work in computers are powered by sequences of voltage drops or machine code, and the human nervous system is one of the most energy-consuming structures within our body. Research is exploring electronic systems that work on the principles of a biological reflex arc, building relatively simple AI systems capable of complex maneuvering.

–R–

Please forward to anyone interested in Artificial Intelligence. If you are new to Radical Reads, sign up here to keep up with the latest trends in AI.