Deep Learning can not be trusted, says Brain Modeling Pioneer

Over the past 20 years, deep learning has come to dominate artificial intelligence research and applications through a range of useful commercial applications. But during the glare are some deep-rooted problems that threaten the rise of technology.

The inability of a typical deep learning program to perform well on more than one task, for example, greatly limits the application of technology to specific tasks in strictly controlled environments. More seriously, it has been argued that deep learning is unbelievable because it cannot be explained – and unsuitable for some applications because it can experience catastrophic forgetfulness. Put more clearly, if the algorithm works, it may be impossible to fully understand why. And while the tool is slowly learning a new database, any part of its learned memories may suddenly collapse. Therefore, it can be risky to apply deep learning to any life-or-death application, such as a medical one.

Now, in a new book, IEEE Fellow Stephen Grossberg argues that a completely different approach is needed. Conscious mind, resonant brain: How each brain makes a mind describes an alternative model for both biological and artificial intelligence based on cognitive and neural research that Grossberg has conducted for decades. He calls his model Adaptive Resonance Theory (ART).

Grossberg – a gifted professor of cognitive and neural systems and of mathematics and statistics, psychology and brain science and biomedical engineering at Boston University – ART based his theories on how the brain processes information.

“Our brains learn to recognize and predict objects and events in a changing world filled with unexpected events,” he says.

Based on this dynamic, ART uses supervised and unsupervised learning methods to solve problems such as pattern recognition and prediction. Algorithms using the theory have been included in major applications such as sonar and radar signal classification, sleep apnea detection, recommended movies, and computer vision-based driver assistance software.

ART can be used with confidence because it is explicable and does not experience catastrophic forgetfulness, says Grossberg. He adds that ART solves what he has called stability-plasticity dilemma: How a brain or other learning system can learn quickly autonomously (plasticity) without experiencing catastrophic forgetfulness (stability).

An illustration of a brain over a blue and red checkered pattern.

Grossberg, who formulated ART in 1976, is a pioneer in modeling how brains become intelligent. He is the founder and director of Boston University’s Center for Adaptive Systems and the founding director of the Center of Excellence for Learning in Education, Science and Technology. Both centers have sought to understand how the brain adapts and learns, and to develop technological applications based on their results.

For Grossberg’s “contribution to understanding brain cognition and behavior and their emulation with technology,” he received the 2017 IEEE Frank Rosenblatt Award, named after the Cornell professor who is considered by some to be “the father of deep learning.”

In his nearly 800-page book, Grossberg attempts to explain how “the little lump of flesh we call a brain” gives rise to thoughts, feelings, hopes, sensations, and plans. In particular, he describes biological neural models that try to explain how it happens. The book also covers the underlying causes of conditions such as Alzheimer’s disease, autism, memory loss and post-traumatic stress disorder.

“Understanding how brains give rise to minds is also important for designing smart systems in computer science, engineering and technology, including AI and smart robots,” he writes. “Many companies have used biologically inspired algorithms of the kind that this book summarizes in several technical and technological applications.”

The theories in the book, he says, are not only useful for understanding the brain, but can also be applied to the design of intelligent systems capable of autonomously adapting to a changing world. Taken together, the book describes the basic process that enables people to be intelligent, autonomous, and versatile.


Grossberg writes that the brain evolved to adapt to new challenges. There is a common set of brain mechanisms that control how people retain information without forgetting what they have already learned, he says.

“We preserve stable memories of past experiences, and these sequences of events are stored in our working memories to help predict our future behaviors,” he says. “People have the ability to continue learning all their lives without new learning washing away memories of important information that we have learned before.”

Understanding how brains give rise to minds is also important for designing smart systems in computer science, engineering and technology, including AI and smart robots.

One of the problems with classical artificial intelligence, he says, is that it often built its models on how the brain power work by using concepts and operations that could be deduced from introspection and common sense.

“Such an approach presupposes that you can introspect inner states of the brain with concepts and words that people use to describe objects and actions in their daily lives,” he writes. “It’s an appealing approach, but its results were all too often insufficient to build a model for how the biological brain really works.”

The problem with today’s artificial intelligence, he says, is that it seeks to mimic the results of brain therapy instead of probing the mechanisms that yield the results. People’s behaviors adapt to new situations and sensations “on the go,” says Grossberg, thanks to specialized circuits in the brain. People can learn from new situations, he adds, and unexpected events are integrated into their accumulated knowledge and expectations about the world.

ART’s network is derived from thought experiments on how humans and animals interact with their environment, he adds. “ART circuits are emerging as computational solutions of several environmental constraints that humans and other terrestrial animals have successfully adapted to ….” This fact suggests that ART design in one form or another may be incorporated into all future autonomous adaptive intelligent devices, whether biological or artificial.

“The future of technology and artificial intelligence will increasingly depend on such self-regulatory systems,” Grossberg concludes. “This is already happening with initiatives such as designing autonomous cars and aircraft. It’s exciting to think about how much more can be achieved when deeper insights into brain design are incorporated into highly funded industrial research and applications. “

From your site’s articles

Related Articles Online


Give a Comment