+- Scivillage.com Casual Discussion Science Forum (https://www.scivillage.com)
+-- Forum: Culture (https://www.scivillage.com/forum-49.html)
+--- Forum: Gadgets & Technology (https://www.scivillage.com/forum-83.html)
+--- Thread: Return of the mechanical battery + New AI masters games without being taught rules (/thread-9565.html)
Return of the mechanical battery + New AI masters games without being taught rules - C C - Dec 28, 2020
Have you ever heard of a mechanical battery?
https://medium.com/predict/have-you-ever-heard-of-a-mechanical-battery-250f8912d9fd
EXCERPTS: . . . The electricity grid is in fact a complex system ... To maintain the stability of the system it is necessary that the energy produced by the generators in a given instant balances the energy demand from the loads, creating a balance between production and consumption or compensating for the difference through the use of storage systems power.
The Flywheel is a heavy wheel attached to a rotating shaft so as to smooth out the delivery of power from a motor to a machine. The inertia of the flywheel opposes and moderates fluctuations in the speed of the engine and stores the excess energy for intermittent use. One of the most promising new technologies for the storage and regulation of network quantities consists in the use of flywheel systems, also called Flywheel Energy Storage (FES).
A rotor characterized by great mechanical inertia is inserted in a robust cylindrical container, in which a certain degree of vacuum is maintained in order to reduce noise and aerodynamic friction. This is also achieved thanks to the use of magnetic bearings or superconducting materials. The flywheel uses electrical energy to accelerate or decelerate and the stored kinetic energy is exchanged with the electrical network through a motor/generator integrated into the rotation axis. The amount of energy stored depends on the rotational speed and inertia of the flywheel, i.e. its shape, mass, and the size of the rotor radius... (MORE - details)
How it works: Flywheel storage
DeepMind's New AI Masters Games Without Even Being Taught the Rules
https://spectrum.ieee.org/tech-talk/robotics/artificial-intelligence/deepminds-new-ai-masters-games-without-even-been-taught-the-rules
EXCERPTS: . . . Now comes MuZero, which doesn't even need to be shown the rules of the game. The new system tries first one action, then another, learning what the rules allow, at the same time noticing the rewards that are proffered—in chess, by delivering checkmate; in Pac-Man, by swallowing a yellow dot. It then alters its methods until it hits on a way to win such rewards more readily—that is, it improves its play. Such learning by observation is ideal for any AI that faces problems that can't be specified easily. In the messy real world—apart from the abstract purity of games—such problems abound.
[...] The system takes a fair amount of computing muscle to train, but once trained, it needs so little processing to make its decisions that the entire operation might be managed on a smartphone. “And even the training isn’t so much,” says Schrittwieser. “An Atari game would take 2-3 weeks to train on a single GPU.”
One reason for the lean operation is that MuZero models only those aspects of its environment—in a game or in the world—that matter in the decision-making process. “After all, knowing an umbrella will keep you dry is more useful to know than modeling the pattern of raindrops in the air,” DeepMind notes, in a statement... (MORE - details)