In a recent study published in nature communications, It was submitted A new kind of artificial intelligence (AI) Able to imitate experts in real time and perform tasks never seen before 3D simulation. This type of AI can quickly learn from the human player and other sources, paving the way for better collaboration between humans and machines. AI typically learns from many human representations, but this requires significant time and effort. Researchers looked for a faster and more efficient method, inspired by how people learn by watching and imitating others. The result is ‘ Artificial intelligence that can learn new things From a human companion in real time, a feature that can be useful in many contexts. Scientists, especially Edward HughesThey used a type of learning called “deep reinforcement learning” to train this AI. After training, the AI has shown that it can effectively learn from humans and other AI systems in complex navigation tasks, even if it has never encountered a human before. For example, it was able to move across complex terrain with many obstacles. A key aspect of this method was the use of a special type of machine learning software that made it possible for the AI to learn progressively, from easier tasks to more difficult ones. This helped make the learning process more effective. These findings not only pave the way for a faster way for AI to learn new things, but it could be one too Useful for improving cooperation between people and machines. Experts suggest that this approach could lead to more adaptable and privacy-friendly AI, inspired by how people learn and interact with each other. Moreover, it could be a useful new tool for scientists studying how cultural abilities develop in humans.
“Infuriatingly humble social media buff. Twitter advocate. Writer. Internet nerd.”
