From Chatbots to Cancer: Using Deep Learning to Solve Problems
Whether we realize it or not, deep learning is transforming our lives. Deep learning is a machine learning technique that teaches computers to learn by example. It enables our voice to control devices like phones, tablets, TVs and hands-free speakers. It is also a key technology behind driverless cars. Dr. Ming Li, Canada Research Chair in Bioinformatics, is exploring this growing field to develop both bioinformatics software and chatting robots.
Being able to talk is at the heart of artificial intelligence (AI), as intelligence is closely related to language. Li and his research team are developing a chatting robot, or chatbot, using deep learning as well as many other techniques. A chatbot is a robot chat that imitates human conversations through voice commands, texts, or both. The one being developed by Li and his team is the first working chatbot that successfully uses general contextual information to chat with the user. For instance, Li’s chatbot, called “Doudou,” writes poems based on any topic provided by a user.
Li and his team are also applying AI and deep learning to develop tools to detect short peptides and neoantigens on the surface of a cancer cell using mass spectrometer. Sensitive detection of such peptides may lead to personalized vaccines for controlling cancer. This approach has the potential to help cancer patients in Canada and around the world receive more personalized, tailored care.