Science Stories
Bedartha Goswami’s goal is to build a bridge between machine learning and climate science. It’s not easy: when new methods in machine learning are developed, the ways that they can be applied in the climate sector is often not considered. Goswami is a team player, so his solution has been to put together a group with the interdisciplinary expertise needed for real breakthroughs in climate science.
Read More
Latest Research
When children develop into adults, how they learn changes a lot. While children show a lot of random behaviour, adults perform more goal-directed actions. An influential theory describes these changes as being similar to the behaviour of an optimisation algorithm commonly used in machine learning. This empirical test shows that there are striking similarities but also important differences between human development and machine learning algorithms.
Read More
Debate
Research in machine learning and data science in and from Africa has the potential to play a more significant global role and faces unique challenges. The pan-African network of AIMS (African Institute for Mathematical Sciences) and its postgraduate programmes prepare young Africans to contribute towards this goal.
Read More
Latest Research
Deep learning algorithms are very good at recognizing specific objects (e.g. a dog, a car) within an image (known as image classifiers). But how do they actually do that? Most often the mechanisms underlying an algorithm’s decision remain opaque. What if we could explain any such black-box algorithm intuitively and, by doing so, even learn from it?
Read More
Debate
Artificial intelligence (AI) and democracy have many touchpoints. What is unclear, however, is whether AI will strengthen or weaken democracy in the long run. It is about time that we, as researchers and citizens, get more involved and develop ideas for a digitally competent democracy together.
Read More
Latest Research
The Bayesian formalism can add uncertainty to deep neural networks. But Bayesian deep learning has a reputation as cumbersome and expensive. No longer. Recent results show how to achieve calibrated uncertainty in deep networks efficiently, without affecting their predictive performance.
Read More