Researchers from Blavatnik School of Computer Science at Tel Aviv University led by Professor Yishay Mansour are developing artificial intelligence algorithm in order to equip machines with human emotions. According to this algorithm based on machine learning, the computer programs are taught a sense of “regret” and how to “feel sorry” for their mistakes.
Using the algorithm developed by Professor Mansour, computer applications are not supposed to “feel” regret, soreness or remorse. It is enough if they simply measure the distance between a desired and the actual outcome. Basing on such “regret metrics” computers will be able to optimize experienced emotions, for example: reduce the amount of virtual regret.
This project, closely related to computer science, game theory and computer-alone decision-making, has been launched by Prof. Mansour at the 23rd Annual Conference on Learning Theory. The Google, which is interested in the potential application of such algorithm on online advertising platforms, is willing to contribute in funding the project. The self-learning algorithm that adapts to the situation in real-time would allow servers and routing systems to recognize sudden spikes in traffic or increased attention to an online product and efficiently prioritize server resource requests. Therefore programs like AdSense and AdWords would be able to deal better with internet traffic and emotions expressed by humans during Internet surfing.
Comments
There are 5 comments: