Seeing the Sky: New Quantum Hardware Offers Millions-Fold More Power
Authors/Creators
Contributors
Producer:
Description
Humanity is now at the early beginning of the deep learning revolution in AI, started by NSF
research grant 0835878 to Andrew Ng, Ed Boyden, Yann LeCun and Yang Dan in 2008. (For
details, see werbos.com/Mind.htm.) The Science News story "Probing Human Mind and Future
Infrastructure Systems, October 3, 2008, described that grant, and the new direction in AI which
gave rise to the COPN program -- cognitive optimization and prediction. Werbos, P. J., & Davis,
J. J. (2016). Regular cycles of forward and backward signal propagation in prefrontal cortex and
in consciousness, in Frontiers in systems neuroscience, 10, 97. went on to show how new neural
network designs based on cognitive prediction and cognitive optimization (compatible with
hardware like GPUs used in early work by LeCun and now by nVidia) can replicate and explain
the higher levels of intelligence we see in mammal brains.
But these designs all rested on minimizing a loss or value function at the bottom, which is usually
a highly nonconvex optimization problem. Likewise, the highest resolution technology for
"seeing the sky" (see QuATh and ICI, by Werbos and Hyland, posted by USPTO in the patent for
thermal quantum Annealing, tQuA), leads to such a nonconvex minimization task. Seeing the
sky better is at the core of many urgent challenges, from warning people of attacks from the
sky to better understanding nuclear forces in our galaxy (like dark matter but with high
resolution). That patent displays a new type of quantum computing, radically different from the
quantum Turing machines known to most experts; it uses quantum superposition to solve these
minimization problems, and, unlike the core of the DWave architecture, uses true physical
annealing to optimize N continuous complex variables. Considerable evidence now
demonstrates the generality of the method, which does for quantum learning what deep learning did for Turing machines.
Files
Werbos_Budapest_AI_Summit_2026 (1).pdf
Additional details
Funding
- Institute of Economics
Dates
- Accepted
-
2026-05-07