Deep Learning Is Hitting a Wall

Fast forward to 2022, and not a single radiologist has been replaced. Rather, the consensus view nowadays is that machine learning for radiology is harder than it looks; at least for now, humans and machines complement each other’s strengths.

Deep learning, which is fundamentally a technique for recognizing patterns, is at its best when all we need are rough-ready results, where stakes are low and perfect results optional.

Manipulating symbols has been essential to computer science since the beginning, at least since the pioneer papers of Alan Turing and John von Neumann, and is still the fundamental staple of virtually all software engineering—yet is treated as a dirty word in deep learning.To think that we can simply abandon symbol-manipulation is to suspend disbel... See more

In 2020, Jared Kaplan and his collaborators at OpenAI suggested that there was a set of “scaling laws” for neural network models of language; they found that the more data they fed into their neural networks, the better those networks performed.10 The implication was that we could do better and better AI if we gather more data and apply deep learni... See more

For at least four reasons, hybrid AI, not deep learning alone (nor symbols alone) seems the best way forward:

1. So much of the world’s knowledge, from recipes to history to technology is currently available mainly or only in symbolic form.


2. Deep learning on its own continues to struggle even in domains as orderly as arithmetic. A hybrid system may have

... See more

To think that we can simply abandon symbol-manipulation is to suspend disbelief.

There are serious holes in the scaling argument. To begin with, the measures that have scaled have not captured what we desperately need to improve: genuine comprehension.

Such signs should be alarming to the autonomous-driving industry, which has largely banked on scaling, rather than on developing more sophisticated reasoning. If scaling doesn’t get us to safe autonomous driving, tens of billions of dollars of investment in scaling could turn out to be for naught.

Classical computer science, of the sort practiced by Turing and von Neumann and everyone after, manipulates symbols in a fashion that we think of as algebraic, and that’s what’s really at stake.