Allen’s Interest in AI, and Philanthropic Synergy

For its fourth round of Allen Distinguished Investigator awards, the Paul G. Allen Family Foundation seems, once again, to be targeting the realm where the line between science and technology blurs. In the cell domain, the Allen Foundation has spent much of 2014 funding projects to generate animated models of cell functions and examining cell translation operations. Now, it’s spreading that same sort of borderline sci-fi interest into the world of Artificial Intelligence.

This round of three-year grants is recognizing five innovative projects, each taking on one of three areas within the field of artificial intelligence: machine reading, diagram interpretation and reasoning, or spatial and temporal reasoning. The researchers and their projects:

  • Devi Parikh (Virginia Tech), who will focus on simplifying the visual world for machines by leveraging abstract scenes to "teach" them common sense;
  • Maneesh Agrawala (University of California, Berkeley) and Jeffrey Heer (University of Washington), who are developing computational models for interpreting graphic visualizations of data and diagrams;
  • Sebastian Riedel (University College London), who will experiment with a new approach to how machines convert symbolic knowledge into vector form and approximate the behavior of logic through algebraic operations;
  • Ali Farhadi and Hannaneh Hajishirzi (University of Washington), who hope to teach computers to interpret diagrams the same way children are taught to do in school; and
  • Luke Zettlemoyer (University of Washington), who will investigate models that enable a machine to automatically read any text book, extract all the knowledge it contains, and use that information to pass a college-level exam on the subject matter.

Lately, Allen seems to be practicing a form of philanthropy grounded in synergy: as it opens up new institutes, and the Paul Allen Family Foundation announces new rounds of Distinguished Investigators, the image I get in my mind is part ripple effect, part Rube Goldberg machine: It seems like everything it’s setting up is in a position to influence or be influenced by something else.

And in realms like these—brain science, cell biology, artificial intelligence—that’s nothing but a good thing. Slowly, it seems, Allen is assembling a small empire of investigators, data, and state-of-the-art technology, investing millions of dollars in the process. It’s pretty clear that Allen feels strongly that the keys to solving many of humanity’s toughest problems lie within the brain. He’s banking on it.

We like that, because it doesn’t seem that his interest in this realm has any one specific target, and because he has many millions to give. His giving, and his institutes, which seem to keep popping up, definitely have our attention.