- This event has passed.
How can a computer understand what is happening in a video?
22 November, 2017 @ 6:30 pm - 7:30 pm
2017 Milner Award Lecture given by Professor Andrew Zisserman FRS.
How can a computer recognise people and what they are doing and saying in a video stream? The answer is by learning, and learning can take many different forms.
One form is known as ‘strong supervision’: this is when a computer is shown many (thousands) of examples of a person or the action they are doing, and from this it learns a model to classify the video. Another form of learning is known as ‘weak’ or ‘self-supervision’: this is when the computer learns directly from the structure of a video stream.
This lecture will explain how both forms of supervision can be used to train neural networks using deep learning. It will be illustrated throughout with examples including: recognising people by their faces, recognising human actions, automated lip reading, and using both sound and images in concord for training.
The prize lecture will be webcast live and the video recording of the event will be available shortly after the event.