Warwick University Business School Researchers have managed to develop a new form of Artificial Intelligence (AI) that deploys a deep learning model to distinguish between the picture of a beautiful scenery and a more run of the mill landscape. In their paper titled, Using deep learning to quantify the beauty of outdoor places, the team put to use 200000 pictures from their website Scenic or Not.
These pictures were fed to the deep learning model to enable it to understand the characteristics of a beautiful picture. The next stage involved the researchers training the deep learning model to rate the pictures according to their beauty. The experiment was so successful that the deep learning model could actually identify Hampstead Heath, a park in London and the iconic Big Ben as a built up area.
What is quite intriguing about this model is the fact that this deep learning model comprised of neural network that is quite similar to the human neural network. The insights provided by this AI model about the collective understanding of human beings about concepts like what is beautiful is fascinating.
What is also incredible is that people who lived in the areas identified as beautiful in this experiment reported better health than the people who loved elsewhere. Another important insight gathered was the nuanced approach to beauty that the AI model took. For instance, just the existence of vast green patches did not constitute beauty, but the presence of architectural features like bridges and structures of historical importance added to the beauty of a place.
This research by Warwick University students surely brought into sharp focus that AI or Artificial Intelligence continues to make vast strides and by and by it is extending its reach into more everyday matters, rather than being confined to only high-end esoteric matters.