Developments over the last decade in big data tools such as deep learning have expanded their use to the point where they take up a large part of the data processing space. However, these methods are based on the availability of a large amount of reliable and unbiased data. So what happens when we have little data and lots of fields?
In this talk, we will look at how in some fields like AI and robotics you can learn with little data. Methods such as reinforcement learning, dimensionality reduction and data representations that do not discard information are crucial to getting good models.
