ABSTRACT:
Autonomous machine learning has become a top priority in science and engineering of learning.
In July 2007, NSF had a workshop on the “Future Challenges for the Science and Engineering of
Learning.” Here is the summary of the “Open Questions in Both Biological and Machine Learning”
from the workshop
(http://www.cnl.salk.edu/Media/NSFWorkshopReport.v4.pdf).
“Biological learners have the ability to learn autonomously, in an ever changing and uncertain
world. This property includes the ability to generate their own supervision, select the most
informative training samples, produce their own loss function, and evaluate their own
performance. More importantly, it appears that biological learners can effectively produce
appropriate internal representations for composable percepts -- a kind of organizational scaffold -
- as part of the learning process. By contrast, virtually all current approaches to machine learning
typically require a human supervisor to design the learning architecture, select the training
examples, design the form of the representation of the training examples, choose the learning
algorithm, set the learning parameters, decide when to stop learning, and choose the way in which
the performance of the learning algorithm is evaluated. This strong dependence on human
supervision is greatly retarding the development and ubiquitous deployment autonomous artificial
learning systems. Although we are beginning to understand some of the learning systems used by
brains, many aspects of autonomous learning have not yet been identified.”
This dismal NSF characterization of the state of our learning systems opens the door to creating a
new generation of learning algorithms. And ICMLA and sister conferences could become the
focal point for research collaboration on this new breed of learning algorithms.
The objective of this tutorial is to present some new ideas regarding brain-like learning, ideas that
can lead to the development of autonomous learning methods. Autonomous learning is extremely
important for robotics. For autonomous robots that can learn on their own like humans, we have
to have tweak-free learning algorithms that can design and train computational structures (e.g.
neural networks) on their own without any kind of human intervention.
Structure of the tutorial:
· Provide an overview of a broad set of principles for designing and constructing autonomous
learning algorithms. Present some new ideas about brain-like learning that differ from current
connectionist approaches.
· Discuss one particular autonomous learning algorithm for pattern classification problems. Give a
demonstration of this autonomous learning algorithm. Summarize its basic features and design
principles.
· As noted in the NSF report, autonomous learning is the technology we need and it is important
that we get organized and focus on this new breed of learning algorithms. So there will be some
open discussion on this issue. We would take this opportunity to form a research group for
collaboration on autonomous learning systems.