In 1985, Terry Sejnowski, a former PhD student of John Hopfield, and Geoffrey Hinton developed the Boltzmann machine, named after physicist Ludwig Boltzmann. Building on Hopfield's energy-based approach to neural networks, it introduced hidden nodes that perform computation between input and output layers, solving a key limitation in training multi-layer neural networks that underlie modern AI systems.
- 01
Sejnowski was at Johns Hopkins and Hinton was at Carnegie Mellon when they developed the model.
- 02
The Boltzmann machine is described as stochastic and recurrent, using an energy-based learning approach similar to Hopfield's earlier work.
279 words · auto-generated from the episode video
1:20:41junction. It established the foundation of superconducting cubits and a lot of quantum technology and fundamentally it's just a very big discovery because >> it shows that the system can be extremely large and still obey quantum mechanics >> which is fascinating and one of our most popular episodes. So be sure to check that out. staying in 1985 invention of the Boltzman machine. >> Yes, this is the next step from the Hopfield network. Terry Sinowski who was actually a PhD student of um John Hopfield, he was at Johns Hopkins at the type time and Jeffrey Hinton, the godfather of AI, he was at Carnegie Melon at the time. They developed the Boltzman machine named after Ludwig van
1:21:22Boltzman, the physicist from 1800s. Um, it's a stochastic recurrent neural network that introduced energy based learning, the same type of energy argument that John Hopfield had used earlier. And it showed how you could learn stuff by having hidden nodes as in middle parts of the network that didn't see the input or the output. Right? There's there's stuff in the middle that is now doing computation. It solved a critical limitation in training multi-layer neural networks which are now the bread and butter of AI today. Big shout out to Jeffrey Hinton there. 1986. Uh again, we're continuing in the same vein, formulation of the back propagation algorithm, a fundamental in
1:22:03all of these chat bots and frontier models we see today. >> This is and we're starting to see why Jeffrey Hinton is called the godfather of AI. The year before he did the Boltzman machine. Now this year with David Rumhalt and Ronald Williams he