2019 outlook: AI research to real-world application
In the first of 6 posts about the Future Labs stage at The AI Summit, New York, we explore key takeaways from industry experts, academic pioneers, and enterprise leaders discussing the next 12–24 months of AI
Artificial intelligence is the rapidly changing present and large future in the business world. That future, with its opportunities and threats, looks different for every company — and business leaders need to understand the foundational research and reality of AI if they want to design the best pathway forward.
The Future Labs — living at the nexus of a leading university and the second largest tech hub in the world, incubating startups commercializing frontier technology, and fostering deep collaboration between innovation, academia, and industry to further AI entrepreneurship — are well placed to bring together insights into AI in 2019 and beyond. On December 5–6, we served as the research partner for The AI Summit, New York, programming a day-long track focused on AI’s ‘Research to Real World’ opportunities, with talks by a 16-strong lineup of industry experts, academic pioneers, and enterprise leaders. Their content covered four key research areas for AI in business (Voice, Vision, Robotics, and Language), plus a look into applications and academic initiatives shaping innovations in each area.
Insights across industries
Our goal was to show how research in AI translates into high-impact use cases, reshaping the way companies work. These industry-specific areas where AI is reaching unprecedented speed and scale include:
Healthcare: Learning image reconstruction AI is drastically accelerating MRI capabilities.
Daniel Sodickson, of NYU Medical School, outlined how AI powers FastMRI technology capable of much quicker imaging and processing times (from 30–60 minutes down to 60–90 seconds) and more natural-looking output images.
The aim goes beyond simply imaging anatomy. Daniel and others are working to ensure that AI learns the physiology behind the phenomena measured through an MRI — helping doctors better predict and diagnose health problems.
Cybersecurity: Enterprises are using data science to capture and analyze more (and more valuable) sources of information to help predict or prevent attacks.
UpLevel Security co-founder and CEO Liz Maida explained how graph modeling helps security analysts understand relationships inside their data, based on similarities of elements like email sender domains, text structure, time/relay, and more.
With the right data infrastructure, algorithms can then elucidate “degrees of connectivity” among an organization’s entire security log — minus hours and hours of search queries it typically takes to contextualize a single incident.
Transportation: Self-driving cars are poised to be the first AI-powered robots with which the general population interacts.
Luc Vincent, VP of Autonomous Driving at Lyft, outlined how his team leverages AI across each of the “building blocks” of a self-driving vehicle — spanning motion control, motion planning, perception, localization, and mapping functionalities — and uses its robotics platform to compute tremendous volume of data use downstream use (like calling a car in its ride-sharing app).
Lyft is one of few companies pursuing Level 5 autonomy, and is collaborating with a number of car companies and other third parties to get there. The application is advanced enough that Lyft is piloting a program to shuttle employees to and from its Palo Alto office every day using self-driving vehicles.
Connecting AI to commerce
For each of these examples, reaping the benefits of AI requires rethinking traditional infrastructures, from how machines ‘see’ to how data is organized and managed. A key theme of our track was the importance of thinking holistically about AI rather than shoehorning it into an existing customer experience or tech stack.
Lyft’s leadership also speaks to how important cross-industry partnerships and platforms are to helping AI reach its full potential. In nearly every area we covered at The AI Summit, our speakers acknowledged the need for deeper collaboration and standard-setting as AI’s commercial applications expand.
So where do we see AI progress? In addition to the above examples, our experts dove into recent innovations in AI that apply for most companies, regardless of industry — including HR/recruiting, IoT, back-office finance, price-monitoring, and many other applications. Next week, we’ll go deeper into the ideas and use cases our speakers explored in our focus areas: Voice, Vision, Robotics, Language, and the Academic/Research Ecosystem. Stay tuned.