Maths4DL: Unlocking the next generation of deep learning

We join Maths4DL researchers for an intensive hackathon working on some of the most challenging problems posed by artificial intelligence.

To make informed decisions regarding AI politicians need to talk to the mathematicians and scientists who develop it. In this episode of Maths on the Move we report on Evidence Week, an event in Parliament that enabled mathematicians and policy makers to exchange ideas.

The type of AI currently in use is based purely on data. How much more powerful would it be if it also has access to humanity's vast knowledge of physics? Find out more in this collection of content.

Combining AI with human knowledge of physics may lead to powerful applications in a range of areas — from weather forecasting to engineering.

Yolanne Lee, A PhD student for the Maths4DL research project, tells us about what she thinks AI will be able to do in the near future, what it has to do with cats and dogs, and how music provided her first experience of science.

Will AI one day be able to spot the equation that describes a physical process just by looking at a video of it? 

If you've flipped heads 10 times what's the next flip most likely to be?  Bayes' theorem has the answer, not just for the coin, but for the pursuit of science generally.

Trying to make a prediction about the world based on dodgy data? Then data assimilation has the answer!

From tiling bathrooms to fooling cancer cells, and from new insights in topology to bringing research into the classroom — we hope you enjoy our April round-up!

Join us at the mathematical frontier with Georg Maierhofer as he tells us about an exciting new idea bringing together physics and machine learning!