Radical Reads — October 13, 2020

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

Radical Reads — October 13, 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) Developments in AI: State of AI Report 2020 (Nathan Benaich and Ian Hogarth)

“We believe that AI will be a force multiplier on technological progress in our increasingly digital, data-driven world. This is because everything around us today, ranging from culture to consumer products, is a product of intelligence.”

Radical Commentary: This week saw the release of the State of AI Report 2020. The Report surveys AI developments across research, talent, industry and politics, along with predictions for the technology over the next 12 months. The 177-slide report emphasizes the importance of AI as a massively disruptive technology across the globe.

A few key observations from this report include:

  • Global demand for AI talent continues to outpace supply, and North America continues to attract most of the global talent pool.
  • AI compute platforms will be essential for unlocking many compute heavy sectors, such as autonomous driving, and both ‘Big Tech’ and privately-backed companies (including Radical portfolio company Untether AI) continue to race each other to release more complex processors every year.
  • Enterprises are continuing to see material benefits in deploying AI applications across all of their major functions.
  • Big Tech, including their affiliates and subsidiaries, continue to invest heavily in AI research and elite talent, and providing them with compute infrastructure and datasets to build more accurate models — a flywheel for success in this sector.
  • Training AI systems is becoming more efficient over time, and efficiency rates have accelerated with advancements in deep learning methods over the past eight years.
  • Regulators and politicians are getting up to speed with AI, and will dive deeper into pressing issues on society, such as facial recognition, deepfakes and algorithmic decision-making.

2) AI and Biology: Deep Learning Takes on Synthetic Biology (Wyss Institute)

“DNA and RNA have been compared to “instruction manuals” containing the information needed for living “machines” to operate. But while electronic machines like computers and robots are designed from the ground up to serve a specific purpose, biological organisms are governed by a much messier, more complex set of functions that lack the predictability of binary code. Inventing new solutions to biological problems requires teasing apart seemingly intractable variables — a task that is daunting to even the most intrepid human brains.

Two teams of scientists from the Wyss Institute at Harvard University and the Massachusetts Institute of Technology have devised pathways around this roadblock by going beyond human brains; they developed a set of machine learning algorithms that can analyze reams of RNA-based “toehold” sequences and predict which ones will be most effective at sensing and responding to a desired target sequence. As reported in two papers published concurrently today in Nature Communications, the algorithms could be generalizable to other problems in synthetic biology as well, and could accelerate the development of biotechnology tools to improve science and medicine and help save lives.”

Radical Commentary: The power of interdisciplinary approaches — melding data science with synthetic biology — helps create new understandings of complex biological systems. Using models from computer vision and natural language processing can help predict better structures of RNA strands (see here for full paper).

Two trends are converging here: (1) the use of deep learning to understand vast volumes of data and predict biologically relevant interactions; and, (2) the “programming” of synthetic biological organisms.These technologies will have a wide range of applications and enable the engineering of cells for diagnostic or therapeutic devices, the production of novel molecules, and even for use in environmental remediation. AI is ushering in a new era of scientific discovery that will accelerate innovation in all fields.

3) AI in Cybersecurity: Will We Have Cyberwar or Cyber Peace? (Wall Street Journal)

“Cyberspace in 2030 could be a very different place than it is today, for good or ill. How we deploy artificial intelligence and machine learning to attack and to defend networks will make the difference.

…AI has the potential to make cyberspace safer for humans. Think of it as a master cyber AI that goes on the defense. Machine learning holds out the theoretical possibility of humans yielding control of network security management, indeed all network operations, to adaptive algorithms. Thus far, however, machine-learning techniques and narrow AI systems have only been incorporated into anomalous activity detection, fraud prevention, and identity and access management tools.”

Radical Commentary: In 2020, there has been an increasing number of cyber attacks. Just this week we saw a crippling incident where hackers attacked the United Nations shipping regulation arm, The International Maritime Organization (IMO).

AI can drastically improve our ability to detect anomalies, reduce the scope of an attack, and bridge the growing talent gap. However, applying AI to the security sector requires particular attention to the security of training data and machine learning models. Security is adversarial in nature and new defense tools or tactics will become a target for malicious actors.

These targets will inevitably include AI approaches. For example, data poisoning and adversarial inputs are means of attacking an AI security management tool. Constant innovation is required to anticipate nefarious use of AI by attackers — including on the AI itself. The majority of attacks today leverage human error and gain access to a system through individuals who are part of an organization and permitted in a system. An entirely AI-based system would need to reconsider the human element in how we use and access systems.

4) AI and Healthcare: The state of artificial intelligence-based FDA-approved medical devices and algorithms: an online database (Nature)

“We provide an insight into the currently available AI/ML-based medical devices and algorithms that have been approved by the US Food & Drugs Administration (FDA). We aimed to raise awareness of the importance of regulatory bodies, clearly stating whether a medical device is AI/ML based or not. Cross-checking and validating all approvals, we identified 64 AI/ML based, FDA approved medical devices and algorithms…

We have launched the first comprehensive and open access database of strictly AI/ML-based medical technologies that have been approved by the FDA. The database will be constantly updated.”

Radical Commentary: There are many obstacles to implementing AI in daily clinical practice, especially in the regulation of these technologies. This database of FDA-approved medical devices and algorithms illustrates that AI in medical devices is beyond the novelty stage. Many of the breakthroughs are in practices with patterns as described by American cardiologist and scientist Dr. Eric Topol with a large majority in radiology-based use cases. Others include cardiology, neurology and general practice/internal medicine.

The commitment to update the database makes it a useful reference for those interested in this growing field.

5) AI and Music Listening: How AI is playing a bigger role in music streaming than you ever imagined (Fortune)

“Behind the scenes of some of the most popular music-streaming services, artificial intelligence is hard at work like an automated DJ, deciding which songs listeners will enjoy.

The technology’s ability to learn from the listening habits of millions of users across millions of songs has made the software key for nearly every music-streaming service today.

But its job doesn’t stop there. A.I. is playing an increasing role in some of the more subtle challenges inherent in music streaming, like adjusting sound volumes and eliminating dead air.”

Radical commentary: Curating a playlist is personal and subjective, but most major music services now use AI to make the experience more enjoyable. AI is playing you the right song, and also ensuring smooth transitions between songs, adjusting sound volumes, and eliminating dead air. The algorithms are also adding surprises into people’s personalized playlists so that they are not listening to the same song too often.

Radical’s founders Jordan Jacobs and Tomi Poutanen’s startup, Milq, pioneered algorithmic personalization in music and then across cultural content, starting in 2011. Theirs was the first service to use deep learning in recommendation systems, delivering the ‘next right song’ (or video) based on a user’s personal taste. That technology was subsequently spun out of Milq as Layer 6 AI, and enabled Layer 6 to defend its 2017 RecSys Challenge title in 2018 when the objective was to deliver the best personalized music playlists via Spotify, ensuring continued listening.

Since then similar systems have been deployed across many item recommendation services ranging from content on Spotify, Apple Music, Netflix, and YouTube, to goods on Amazon.


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) America’s Racial Gap & Big Tech’s Closing Window (Deutsche Bank Research)

“The exponential growth of the digital economy is going to leave large chunks of minorities with little or no access to jobs. We conduct a bottom up societal study and it shows that 76% of Blacks and 62% of Hispanics could get shut out or be under-prepared for 86% of jobs in the US by 2045. If this digital racial gap is not addressed, in one generation alone, digitization could render the country’s minorities into an unemployment abyss.

We went into the study expecting a gap, but the data is far more glaring. Due to the structural and infrastructural inequities, Blacks and Hispanics are 10 years behind Whites in levels of broadband access and almost 4 times more Blacks have poor Tech connectivity than Whites.”

— R —