Machine learning systems maybe be smart but they have a lot to discover.
Innovative researchers with DARPA hope to achieve superior machine learning systems with a new program called Lifelong Learning Machines (L2M) which has as its primary goal to develop next-generation machine learning technologies that can learn from new situations and apply that learning to become better and more reliable than current constrained systems.
+More on Network World: DARPA fortifies early warning system for power-grid cyber assault+
“Life is by definition unpredictable. It is impossible for programmers to anticipate every problematic or surprising situation that might arise, which means existing ML systems remain susceptible to failures as they encounter the irregularities and unpredictability of real-world circumstances,” said L2M program manager Hava Siegelmann in a statement. “Today, if you want to extend a machine learning system’s ability to perform in a new kind of situation, you have to take the system out of service and retrain it with additional data sets relevant to that new situation. This approach is just not scalable.”
Machine learning methods have demonstrated outstanding recent progress; as a result, AI systems can now be found in myriad applications, including autonomous vehicles, industrial applications, search engines, computer gaming, health record automation, and big data analysis. However, today’s AI can only operate in very orchestrated, specific environments with an extensive training set that exactingly describes the conditions that will occur during execution time, DARPA stated.
+More on Network World: Gartner: Algorithm-based technology patents are raging+
The four-year L2M program aims to change all that by focusing development on two technical areas. The first technical area will explore algorithms, theoretical modeling, analysis, software, and architectures that implement new approaches for continuous learning. It will also pursue robustness and safety by setting limits on system behaviors and allowing users both to monitor the system’s behavior and evolution and intervene as needed, DARPA stated. The idea is to continuously apply the results of experience and adapt “lessons learned” to new data or situations. Simultaneously, it calls for the development of techniques for monitoring a machine learning system’s behavior, setting limits on the scope of its ability to adapt, and intervening in the system’s functions as needed, DARPA said.
The second technical will focus specifically on how living systems learn and adapt and will consider whether and how those principles and techniques can be applied to machine learning systems. Here what DARPA called novel algorithms will be validated by implementing the approach as either a component of the first technical area lifelong learning system or other proposed applications. Research within technical area 2 is likely to require integration of the field of computer science with other domain areas of nature, e.g., biology and chemistry. Concepts from nature could include but are not limited to: Mechanisms for evolving networks; memory stability in adaptive networks; goal-driven behavior mechanisms and learning rules and plasticity mechanisms.
“Enabling a computer to learn even the simplest things from experience has been a longstanding but elusive goal,” said Siegelmann. “That’s because today’s computers are designed to run on prewritten programs incapable of adapting as they execute, a model that hasn’t changed since the British polymath Alan Turing developed the earliest computing machines in the 1930s. L2M calls for a new computing paradigm.”
Check out these other hot stories: