Reinforcement learning is a continuously growing field of AI where each day new application of AI pops up, this is one of them. Recent research done by Deepmind shows the AI agent can navigate to cities without a map. We all know self-driving cars need a map to navigate to the destination but with this recent research, AI can navigate with the help of navigation instruction. This is similar to navigating in games.

This AI agent is called StreetNav which takes help of Google Street View and provides visually accurate environments representing real places. This artificial agent is a combination of neural network, RL, Computer vision.

Deepmind in this blog (https://deepmind.com/blog/learning-to-navigate-cities-without-a-map/) said, “Studying navigation is fundamental in the study and development of artificial intelligence, and trying to replicate navigation in artificial agents can also help scientists understand its biological underpinnings.”

You can read more about it here.[https://arxiv.org/abs/1903.00401]

Leave a Reply