More on this book
Community
Kindle Notes & Highlights
by
Cade Metz
Started reading
September 18, 2022
neuroscientist at Joh...
This highlight has been truncated due to consecutive passage length restrictions.
“the neural network und...
This highlight has been truncated due to consecutive passage length restrictions.
Hinton and Sejnowski,
believed the future still lay in systems that could learn behavior on their own.
when it came to neural networks, Minsky might have attacked them as woefully deficient—
Hinton saw him as a “lapsed neural-net-er,”
Hinton was born in Wimbledon, England,
It was expected that he would one day follow his father into academia.
It was less clear what he would study.
He wanted to study t...
This highlight has been truncated due to consecutive passage length restrictions.
As an undergraduate at King’s College, Cambridge, what he wanted was a better understanding of the brain.
no one understood much more than he did.
Hinton tried physiology and chemistry, physics and psychology.
In the end,
Hinton left academia entirely.
After graduating from Cambridge,
Hinton moved to London and became a carpenter.
That year, he read The Organization of Behavior,
to explain the basic biological process that allowed the brain to learn.
Learning, Hebb believed, was the result of tiny electrical signals that fired along a series of...
This highlight has been truncated due to consecutive passage length restrictions.
change that wired these neurons together...
This highlight has been truncated due to consecutive passage length restrictions.
“Neurons that fire together, wir...
This highlight has been truncated due to consecutive passage length restrictions.
theory—known as Hebb’s Law—helped inspire the artificial neural networks built by scient...
This highlight has been truncated due to consecutive passage length restrictions.
like Frank Rosenblatt before him, Hinton came to believe that each side—the biological and the artificial—could help the other move forward.
to the AI program in Edinburgh.
Edinburgh, he won a spot in a lab overseen by
Christopher Longuet-Higgins.
he abandoned brainlike architectures and switched to symbolic AI—
Hinton spent his graduate school years
dismissed not only by his colleagues but by h...
This highlight has been truncated due to consecutive passage length restrictions.
wasn’t all that interested in mathematics,
including the linear algebra that drove neural networks.
Minsky and Papert’s book pushed most researchers away from connectionism, it drew Hinton closer.
Minsky and Papert, he felt,
almost a caricature of Rosenb...
This highlight has been truncated due to consecutive passage length restrictions.
He had no choice but to look abroad, including the United States.
AI research was on the wane in the U.S., too,
much to his own surprise, he found a small group of people who believed in the same ideas he did.
Crick had won a Nobel Prize for discovering the structure of the DNA molecule
before turning his attention to the brain.
fall of 1979, he published a call to arms in the pages of Scient...
This highlight has been truncated due to consecutive passage length restrictions.
to at least make an attempt at understanding how ...
This highlight has been truncated due to consecutive passage length restrictions.
kind of
Rosenblatt believed that if researchers could build a multilayered network, each layer feeding information into the next, this system could learn the complex patterns his Perceptron could not—
a more brainlike system would emerge.
Give a photo of, say, a dog to this more complex system, and a far more sophisticated analysis would follow.
One of the leading figures in the PDP group was
David Rumelhart,
Rumelhart had set himself a very particular, but central, challenge.
difficult to determine the relative importance (“the weight”) of each neuron to the calculation as a whole.
