Exploration and exploitation with small data

Offentlig prosjektsammendrag

AI research has been driven by large corporations with massive amounts of data. Focus has been on solving big data problems like optimizing websites with millions of views every day. However, this is not the reality in most most companies or government institutions. They don't have big data, they have small data. In this project we will develop AI for sparse data. More specifically in settings where decisions are made after each other, for instance every week, so called sequential decision making. Here you will always have the dilemma of doing what you have always done and know that works fairly well, and something new, with promise, that you hope can do better. This is the type of problem we will help train computers to solve.