Learning of user models in human-in-the-loop machine learning

This project is related to Turing AI World-Leading Researcher Fellowship: Human-AI Research Teams - Steering AI in Experimental Design and Decision-Making.
Machine learning offers great promise in helping us solving problems by automatically learning solutions from data, without us having to specify all details of the solution as in early computational approaches. However will still need to tell machine learning system what problems we want them to solve, and this is currently undertaken by specifying desired outcomes and designing objective functions and rewards. From formulating the rewards for a new problem is not easy for us as humans, and is particularly difficult when we only partially know the goal. As it's the case at the beginning of scientific research. In this programme we developed ways for machine learning systems to help humans to steal them in the process of collecting more information by designing experiments, interpreting what the results mean, and deciding what to measure next, to finally reach the conclusion and Chester with the solution to the problem.
The machine learning techniques will be developed first for three particular important problems and then generalised to be broadly applicable. First is to diagnosis and treatment decision making in personalised the machine, and the second steering of the scientifically experiments in the synthetic biology and drug design, and the third design and the use of digital twins in designing physical systems and processes. And an AI centre of excellence will be established at the University of Manchester, in collaboration with the Turing institute and number of partners from the industry and healthcare sector, and with strong Connexions to the networks best national and international AI researchers.
A brief summary of the project is:
- We aim to build an AI assistant to push the frontier of AI design.
- The AI assistant is cooperative instead of automatic with a free pass.
- We use the algorithm theory to design the mechanism, which focuses on the incentives of humans.
- Three types of agents, the mechanism designer, agents act in the mechanism and an AI assistant, are included in the project.
- The AI assistant helps the mechanism designer to design the mechanism in which the agents act in.
- The AI assistant is inclusive via considering the computational bound of humans instead of perfect rationality.
- The projects would enhance human intelligence with machine intelligence as AI can easily simulate with reasonable accuracy.
The goal is to develop and test Bayesian inference techniques for learning of advanced user models from observational data. The problem setting resembles inverse reinforcement learning, but new techniques need to be developed to cope with the model evolving along time as the user learns, and the user model has several nested multi-agent levels, and bounded-rationality constraints from cognitive science. The student will work with a team of researchers, co-supervised by top-level experts in machine learning Professor Samuel Kaski and Professor Xiaojun Zeng, and be able to apply the techniques in several exciting use cases with industry and academics of other fields.
Professor Sami Kaski from the Department of Computer Science has been appointed among the first Turing Artificial Intelligence (AI) World-Leading Research Fellow. The fellowships, named after AI pioneer Alan Turing, are part of the UKRI's commitment to further strengthen its position as a global leader in the field.
Through his fellowship, Professor Kaski aims to overcome a fundamental limitation of current AI systems, that they require a detailed specification of the goal before they can help. Machine learning, where solutions to problems are automatically learnt from data, is a form of AI with great promise for addressing a number of challenges. This includes healthcare, where AI can detect patterns associated with diseases and health conditions by studying data.