Explainable AI Seminars @ Imperial

CLArg Group, Centre for eXplainable Artificial Intelligence, Imperial College London

XAI has witnessed unprecedented growth in both academia and industry in recent years (alongside AI itself), given its crucial role in supporting human-AI partnerships. XAI allows to deploy opaque data-driven AI methods safely in a variety of settings, such as finance, healthcare and law. XAI is positioned at the intersection of AI, human-computer interaction, the social sciences (in particular psychology) and applications.
Overall, XAI is increasingly part of all AI policies on ethics, trustworthiness and safety of AI. This seminar series focuses on all aspects of XAI, ranging from methods to applications.

Next Seminar

Patrick Altmeyer: "ECCCos from the Black Box: Faithful Model Explanations through Energy-Constrained Conformal Counterfactuals"

13^th May, 16:00 BST, Imperial College London South Kensington Campus, Huxley Building LT308 (directions) and on Microsoft Teams (join online).

Abstract

Counterfactual explanations offer an intuitive and straightforward way to explain black-box models and offer algorithmic recourse to individuals. To address the need for plausible explanations, existing work has primarily relied on surrogate models to learn how the input data is distributed. This effectively reallocates the task of learning realistic explanations for the data from the model itself to the surrogate. Consequently, the generated explanations may seem plausible to humans but need not necessarily describe the behaviour of the black-box model faithfully. We formalise this notion of faithfulness through the introduction of a tailored evaluation metric and propose a novel algorithmic framework for generating Energy-Constrained Conformal Counterfactuals that are only as plausible as the model permits. Through extensive empirical studies, we demonstrate that ECCCo reconciles the need for faithfulness and plausibility. In particular, we show that for models with gradient access, it is possible to achieve state-of-the-art performance without the need for surrogate models. To do so, our framework relies solely on properties defining the black-box model itself by leveraging recent advances in energy-based modelling and conformal prediction. To our knowledge, this is the first venture in this direction for generating faithful counterfactual explanations. Thus, we anticipate that ECCCo can serve as a baseline for future research. We believe that our work opens avenues for researchers and practitioners seeking tools to better distinguish trustworthy from unreliable models.

Bio

Patrick is a PhD Candidate in Trustworthy Artificial Intelligence at Delft University of Technology working on the intersection of Computer Science and Finance. His current research revolves around Counterfactual Explanations and Probabilistic Machine Learning. Previously, Patrick worked as an Economist for the Bank of England. He is also an active contributor and member of the Julia community and have open-sourced a range of packages geared towards Trustworthy AI.

Planned Seminars