Scott Jones

Research Interests

• Third Wave AI

  We are currently experiencing the Second Wave of AI, dominated by deep learning and statistical, "big data" approaches to AI. But just around the corner, perhaps sooner than many people realize, is the Third Wave of AI.

  Third Wave AI systems will feature dramatic improvements, most notably in their ability for contextual adaptation. They will understand context and meaning, and be able to adapt accordingly. Third Wave AI will not only recognize the cat, but will be able to explain why it's a cat and how it arrived at that conclusion – a giant leap from today's "black box" systems.

• Neuro-Symbolic AI

  From the 1970's to the 1990's, the field of AI was driven by heuristic rule-based systems. Known as "Good Old-Fashioned AI" or GOFAI, these systems, which included medical diagnostic applications and chess playing programs, were very good at reasoning, but incapable of learning without reprogramming. By contrast, the last two decades, from the early 2000's to present, have been dominated by statistical machine learning (ML) techniques, which are excellent at learning but have no ability to reason whatsoever. A ML model that's trained to recognize images of cats, for example, can do so with an accuracy level approaching that of humans, yet it has no idea what a cat is. This lack of contextual awareness and inability to explain its results, along with the need for enormous amounts of training data, are major limitations of ML.

  Neuro-Symbolic AI (NSAI) represents an attempt to marry the best of both worlds, combining the reasoning abilities of GOFAI with the powerful, automated learning of ML. NSAI aims to build systems that can learn with substantially less data, and that provide understandable and controllable decisions and actions. Many researchers, including IBM Research, view NSAI as a pathway to Artificial General Intelligence (AGI).

• Artificial General Intelligence (AGI)

  Artificial General Intelligence (AGI) refers to machines that can perform any intellectual task that a human can. Also called “strong AI” or “real AI,” AGI represents the “holy grail” and the original goal of artificial intelligence research since the 1950s.

Current Projects