“Plato” is the name of the platform for intelligent interactive systems developed by Uber artificial intelligence researchers in the US, led by a Greek scientist.
Smart interactive systems have been significantly developed in recent decades, with the most prominent example being the smart personal digital assistants (type “Siri” or “Cortana”), which tend to become essential in everyday life.
The team of researchers at Uber’s Artificial Intelligence Department, led by Alexander Papangelis, developed the platform “Plato Research Dialogue System”, which facilitates the testing, evaluation and improvement of new algorithms, these intelligent interactive systems based on, and the systems themselves.
With “Plato’s” help, researchers can develop and train new dialogue systems (AI agents), even if their users have little experience in the field.
To date, many attempts have been made to create such platforms for general or more specialized use, such as Olympus, PyDial, ParlAI, Virtual Human Toolkit, Rasa, DeepPavlov, ConvLab and more. Most, however, have several disadvantages, such as assuming that users are familiar with the programming code of that platform, the platform is too limited in targeting, or requires a license.
The new flexible platform “Plato”, presented and used by Mr. Papangelis in a lesson he taught in July at a summer school at the University of Crete, overcomes these difficulties and can be used to create, train and evaluate all kinds of smart interactive systems.
It supports machine interactions with people through speech, text or structured information, as well as various dialog architectures.
Mr. Papagelis studied Computer Science at the University of Athens, did his masters degree at the University College London (UCL) and received his Ph.D. from the University of Texas.
He originally worked at Toshiba’s Cambridge Research Center and today is a senior researcher at Uber’s interactive artificial intelligence team in the US.