Learning and deploying safe and trustworthy models of data provenance

Our modern lives are increasingly governed by ubiquitous AI systems and an abundance of digital data. More and more products and services are providing us with better tools and recommendations for our professional, personal, and entertainment activities. With the...

Generative modelling with neural probabilistic circuits

The current state of the art in generative modelling is dominated by neural networks. Despite their impressive performance on many benchmark tasks, these algorithms do not provide tractable inference for common and important queries, e.g. marginalization and...

Dealing with Imperfect Rationality in AI Systems

AI systems often collect their input from humans. For example, parents are asked to input their preferences over primary schools before a centralised algorithm allocates children to schools. Should the AI trust the input provided by parents who may try to game the...

Data Bias Evaluation and Mitigation via Rule-based Classification

Motivation: Training data can be severely biased. The existing metrics of data bias are based on data balance situations conditioned on protected attributes. This is coarse-grained and does not consider the relationship among different attributes as well as different...

Detecting Deception and Manipulation in Planning and Explanation Systems

Planning algorithms are used in a variety of contexts, from navigation apps to recommendation algorithms, robot vacuums, autonomous vehicles, etc.Companies using such algorithms have financial incentives to manipulate (or nudge) user behaviour so as to obtain valuable...

Safe Reinforcement Learning from Human Feedback

Reinforcement learning (RL) has become a new paradigm for solving complex decision-making problems. However, it presents numerous safety concerns in real world decision making, such as unsafe exploration, unrealistic reward function, etc. As reinforcement learning...