Deep learning for Conversational AI
In this talk, I will review recent progress in building conversational AI to complete tasks and hold general conversations. Today's dialogue systems are the fruit of decades of progress in linguistics, compute power, and machine learning. In particular, modern machine learning techniques like deep learning hold the promise of accelerating the development of dialogue systems and achieving more complex interactions but they also entail many challenges including controllability, evaluation, and data efficiency. I will describe these challenges as well as some of the promising solutions that will bring the next generation of dialogue systems.
Layla El Asri is a team lead at Borealis AI. She completed a Ph.D. in computer science at Université de Lorraine in France in 2016. Her Ph.D. was a joint project between Université de Lorraine and Orange Labs, the research and development branch of Orange, a telecommunication company in France. Her research focused on improving dialogue systems with machine learning. She developed methods to train dialogue systems faster while respecting strong industrial constraints. After her Ph.D, she joined Maluuba, a Canadian startup, in 2016 as a research scientist where she worked on user simulation, reinforcement learning, and datasets for dialogue systems. Layla then joined Microsoft, through the acquisition of Maluuba, in 2017 as a research manager leading a team focused on conversational AI and natural language processing. She is continuing her work on natural language processing at Borealis AI.