The T-Bank AI Research laboratory has presented the first open virtual environment for rapid contextual reinforcement learning — XLand-MiniGrid. This solution will allow artificial intelligence (AI) to learn not from scratch, but by interacting with the environment and using context clues.
We came to the field of contextual reinforcement learning when it was still in its infancy, and therefore did not find a single suitable tool for evaluating new ideas. It became clear that this is a problem for many specialists, which means that it should be one of the first to be solved.
XLand-MiniGrid has already attracted the attention of researchers from leading organizations such as Google DeepMind, the University of California, Berkeley, and the University of Oxford.
Vladislav Kurenkov, Head of the Adaptive Agents Research Group at the AIRI Institute, notes that XLand-MiniGrid will not only speed up the process of creating new algorithms, but also open up new horizons for research in the field of artificial intelligence.
Earlier www1.ru reported that in Russia, AI will be trained to create cancer vaccines.
Read materials on the topic:
Yandex has come up with a new way to compress neural networks
A neural network for finding employees on social networks was created in St. Petersburg
Two neural networks have begun testing on the «Gosuslugi» portal