NTT Research Scientist Creates Model for Discontinuous Learning
The Californer/10246919

Trending...
SUNNYVALE, Calif. ~ Dr. Gautam Reddy of NTT Research's Physics & Informatics (PHI) Lab has made a novel contribution to the understanding of learning with his paper, "A reinforcement-based mechanism for discontinuous learning," published in the Proceedings of the National Academy of Sciences (PNAS). The paper uses data on spatial navigation of laboratory mice to provide a simple explanation for the "a-ha" moment of sudden insight.

Dr. Reddy's research shows that the outcome of reinforcement learning (RL) algorithms, which operate gradually, can appear discontinuous for certain kinds of tasks. Specifically, this is true when the artificial RL agent learns the direct path from point A to point B in certain structured environments. This is similar to how humans mentally explore and navigate a knowledge database.

The paper's novel contribution is that RL algorithms can reproduce "a-ha" moment discontinuities in simulations. Dr. Reddy quantifies the relationship between learning rule, derived from biologically based learning discontinuities; learning curve, which represents the speed of reinforcement waves; and exploration bias, or the probability that a mouse will take a certain action at a certain intersection.

More on The Californer
Yoshihisa Yamamoto, Director of PHI Lab said: "Congratulations to Dr. Reddy on his publication in this prestigious journal and for his creative discovery. His important findings will advance our understanding of how we reason, solve problems and gain insight."

Dr. Reddy said: "We show that the nonlinear dynamics of RL-based learning, together with continuous exploration, lead to discontinuous learning curves in tree-like structured environments." He added: "We develop a quantitative theory which explains the origin and highlights the generality of the phenomenon. The theory explains existing data and provides specific testable predictions."

Looking ahead, Dr. Reddy plans to extend his experimental framework beyond tree-like graphs to topologies with loops as well as examine how an animal learns its location in a maze before evaluating different actions at that location. He is also leading NTT Research PHI Lab's five-year joint research project with scientists at Harvard University to study animal neuro-responses and inform future AI systems.

More on The Californer
In addition to Harvard University, eight other universities have agreed to conduct joint research with PHI Lab including Caltech, Cornell University, MIT, Notre Dame University, Stanford University, Swinburne University of Technology Tokyo Institute of Technology and University of Michigan while NASA Ames Research Center has also entered into a joint research agreement with PHI Lab as well.

The mission statement for PHI Lab is to uncover fundamental principles and novel technologies that advance information processing beyond state of the art by exploring quantum information science neuroscience and optical technologies as evidenced by their October 2020 paper in Applied Physics Letters (APL) titled "Coherent Ising Machines: Quantum optics and neural network perspectives" coauthored by Yamamoto himself which underscored their interdisciplinary nature pathbreaking research agenda .
Filed Under: Business
Show All News | Report Violation

0 Comments

Latest on The Californer