Translating Videos to Commands for Robots with Deep Recurrent Neural Networks

Researchers are always looking for new ways to make it easier to teach robots new commands. Anh Nguyen, Dimitrios Kanoulas, Luca Muratore, Darwin G. Caldwell, Nikos and G. Tsagarakis have presented a paper that explores translating videos to commands with deep recurrent neural networks. The framework extracts features from the video and uses two RNN layers with an encoder-decoder architecture to encode the visual features and generate the commands.

[Paper]