Are AI systems limited in their ability to learn/incorporate new information?

This past week Nature magazine published an article at https://doi.org/10.1038/s41586-024-07711-7  titled "Loss of Plasticity in Deep Continual Learning" by Shibhansh Dohare and others.

It basically claims that AI systems can not go on learning new things forever.

In contrast, the Human brain can continually learn and incorporate new information, over the whole of it's "defect free" lifetime.

The paper suggests a possible way of correcting this basic AI learning problem by resetting certain weights used in its neural network. This effectively introduces noise into the system.

Peter Brooks

Palm Bay Florida 

Parents
  • Hi Peter, Can we expect one AI system to be specialist in everything. In the realm of engineering, we often rely on specialists, whether it’s a structural engineer, a Electrical Consultant, or a electrician. Each brings deep expertise in their respective fields, contributing to more efficient and accurate problem solving. Could our AI models follow a similar path?

    - Andrew

  • Hello Andrew:

    It will go wherever there is the most "money" to be made.

    Each day I read numerous "medical" related articles in JAMA and NEJM and they are full of AI related articles some positive and many highly negative.

    Doctors don't like the idea of losing their power and position in society.  

    Peter Brooks

    Palm Bay

Reply
  • Hello Andrew:

    It will go wherever there is the most "money" to be made.

    Each day I read numerous "medical" related articles in JAMA and NEJM and they are full of AI related articles some positive and many highly negative.

    Doctors don't like the idea of losing their power and position in society.  

    Peter Brooks

    Palm Bay

Children
No Data