complex systems

Reinforcement learning of artificial intelligence. Illustration: depositphotos.com

Unleashing the potential of artificial intelligence: a new algorithm developed at MIT improves neural network training efficiency by 50 times

MIT researchers have developed an advanced algorithm for reinforcement learning (initial training of an artificial intelligence system) that optimizes the decision-making of AI systems in complex tasks such as urban traffic control, while significantly saving time and resources
Themis, the blind goddess of justice. Image: depositphotos.com

Probability in morality and law

An unconventional combination of philosophy, law and theory of complex systems, offers surprising insights
Adjusters of the movement velocities of water drops in a two-dimensional system. Red marking represents droplet pairs with a positive correlation, and blue marking represents droplet pairs with a negative correlation

order in disorder