In the past few decades, intelligent conversational agents have undergone significant changes, from keyword-targeted interactive voice response (IVR) systems to cross-platform intelligent personal assistants that are becoming an integral part of everyday life. With this growth, an intuitive, flexible, and comprehensive research and development platform is needed that can serve as an open testing platform to help evaluate new algorithms, quickly prototype, and reliably deploy session agents.
In this context, Uber developed and opened up the Plato research dialogue system. Plato is designed to build, train, and deploy session AI agents, enabling data scientists and amateurs to collect data from prototypes and demonstration systems. It provides a clean and easy to understand design and integrates with existing deep learning and Bayesian optimization frameworks to reduce the need to write code.
“We believe that Plato has the capability to more seamlessly train conversational agents across deep learning frameworks, from Ludwig and TensorFlow to PyTorch, Keras, and other open-source projects, leading to improved conversational AI technologies across academic and industry applications,” wrote Uber AI researchers Alexandros Papangelis, Yi-Chia Wang, Mahdi Namazifar, and Chandra Khatri. “[We’ve] leverage[d] Plato to easily train a conversational agent how to ask for restaurant information and another agent how to provide such information; over time, their conversations become more and more natural.”
Plato was released in open source on GitHub.