Research Papers

The application of Algorithmic Recourse in decision-making is a promising field that offers practical solutions to reverse unfavorable decisions. However, the inability of these methods to consider potential dependencies among variables poses a significant challenge due to the assumption of feature independence. Recent advancements have incorporated knowledge of causal dependencies, thereby enhancing the quality of the recommended recourse actions. Despite these improvements, the inability to incorporate the temporal dimension remains a significant limitation of these approaches. This is particularly problematic as identifying and addressing the root causes of undesired outcomes requires understanding time-dependent relationships between variables. In this work, we motivate the need to integrate the temporal dimension into causal algorithmic recourse methods to enhance recommendations' plausibility and reliability. The experimental evaluation highlights the significance of the role of time in this field.

This paper explores the journey of AI in finance, with a particular focus on the crucial role and potential of Explainable AI (XAI). We trace AI's evolution from early statistical methods to sophisticated machine learning, highlighting XAI's role in popular financial applications. The paper underscores the superior interpretability of methods like Shapley values compared to traditional linear regression in complex financial scenarios. It emphasizes the necessity of further XAI research, given forthcoming EU regulations. The paper demonstrates, through simulations, that XAI enhances trust in AI systems, fostering more responsible decision-making within finance.

Counterfactuals of Counterfactuals: a back-translation-inspired approach to analyse counterfactual editors

Giorgos Filandrianos, Edmund Dervakos, Orfeas Menis-Mastromichalakis +2

In the wake of responsible AI, interpretability methods, which attempt to provide an explanation for the predictions of neural models have seen rapid progress. In this work, we are concerned with explanations that are applicable to natural language processing (NLP) models and tasks, and we focus specifically on the analysis of counterfactual, contrastive explanations. We note that while there have been several explainers proposed to produce counterfactual explanations, their behaviour can vary significantly and the lack of a universal ground truth for the counterfactual edits imposes an insuperable barrier on their evaluation. We propose a new back translation-inspired evaluation methodology that utilises earlier outputs of the explainer as ground truth proxies to investigate the consistency of explainers. We show that by iteratively feeding the counterfactual to the explainer we can obtain valuable insights into the behaviour of both the predictor and the explainer models, and infer patterns that would be otherwise obscured. Using this methodology, we conduct a thorough analysis and propose a novel metric to evaluate the consistency of counterfactual generation approaches with different characteristics across available performance indicators.

EXplainable Artificial Intelligence (XAI) aims to help users to grasp the reasoning behind the predictions of an Artificial Intelligence (AI) system. Many XAI approaches have emerged in recent years. Consequently, a subfield related to the evaluation of XAI methods has gained considerable attention, with the aim to determine which methods provide the best explanation using various approaches and criteria. However, the literature lacks a comparison of the evaluation metrics themselves, that one can use to evaluate XAI methods. This work aims to fill this gap by comparing 14 different metrics when applied to nine state-of-the-art XAI methods and three dummy methods (e.g., random saliency maps) used as references. Experimental results show which of these metrics produces highly correlated results, indicating potential redundancy. We also demonstrate the significant impact of varying the baseline hyperparameter on the evaluation metric values. Finally, we use dummy methods to assess the reliability of metrics in terms of ranking, pointing out their limitations.

Pittsburgh Learning Classifier Systems for Explainable Reinforcement Learning: Comparing with XCS

Jordan T. Bishop, Marcus Gallagher, Will N. Browne

Interest in reinforcement learning (RL) has recently surged due to the application of deep learning techniques, but these connectionist approaches are opaque compared with symbolic systems. Learning Classifier Systems (LCSs) are evolutionary machine learning systems that can be categorised as eXplainable AI (XAI) due to their rule-based nature. Michigan LCSs are commonly used in RL domains as the alternative Pittsburgh systems (e.g. SAMUEL) suffer from complex algorithmic design and high computational requirements; however they can produce more compact/interpretable solutions than Michigan systems. We aim to develop two novel Pittsburgh LCSs to address RL domains: PPL-DL and PPL-ST. The former acts as a "zeroth-level" system, and the latter revisits SAMUEL's core Monte Carlo learning mechanism for estimating rule strength. We compare our two Pittsburgh systems to the Michigan system XCS across deterministic and stochastic FrozenLake environments. Results show that PPL-ST performs on-par or better than PPL-DL and outperforms XCS in the presence of high levels of environmental uncertainty. Rulesets evolved by PPL-ST can achieve higher performance than those evolved by XCS, but in a more parsimonious and therefore more interpretable fashion, albeit with higher computational cost. This indicates that PPL-ST is an LCS well-suited to producing explainable policies in RL domains.

Explainable artificial intelligence (AI) techniques are increasingly being explored to provide insights into why AI and machine learning (ML) models provide a certain outcome in various applications. However, there has been limited exploration of explainable AI techniques on time-series data, especially in the healthcare context. In this paper, we describe a threshold-based method that utilizes a weakly supervised model and a gradient-based explainable AI technique (i.e. saliency map) and explore its feasibility to identify salient frames of time-series data. Using the dataset from 15 post-stroke survivors performing three upper-limb exercises and labels on whether a compensatory motion is observed or not, we implemented a feed-forward neural network model and utilized gradients of each input on model outcomes to identify salient frames that involve compensatory motions. According to the evaluation using frame-level annotations, our approach achieved a recall of 0.96 and an F2-score of 0.91. Our results demonstrated the potential of a gradient-based explainable AI technique (e.g. saliency map) for time-series data, such as highlighting the frames of a video that therapists should focus on reviewing and reducing the efforts on frame-level labeling for model training.

As Machine Learning models are considered for autonomous decisions with significant social impact, the need for understanding how these models work rises rapidly. Explainable Artificial Intelligence (XAI) aims to provide interpretations for predictions made by Machine Learning models, in order to make the model trustworthy and more transparent for the user. For example, selecting relevant input variables for the problem directly impacts the model's ability to learn and make accurate predictions, so obtaining information about input importance play a crucial role when training the model. One of the main XAI techniques to obtain input variable importance is the sensitivity analysis based on partial derivatives. However, existing literature of this method provide no justification of the aggregation metrics used to retrieved information from the partial derivatives. In this paper, a theoretical framework is proposed to study sensitivities of ML models using metric techniques. From this metric interpretation, a complete family of new quantitative metrics called $α$-curves is extracted. These $α$-curves provide information with greater depth on the importance of the input variables for a machine learning model than existing XAI methods in the literature. We demonstrate the effectiveness of the $α$-curves using synthetic and real datasets, comparing the results against other XAI methods for variable importance and validating the analysis results with the ground truth or literature information.

Bridging the Gap: Gaze Events as Interpretable Concepts to Explain Deep Neural Sequence Models

Daniel G. Krakowczyk, Paul Prasse, David R. Reich +3

Recent work in XAI for eye tracking data has evaluated the suitability of feature attribution methods to explain the output of deep neural sequence models for the task of oculomotric biometric identification. These methods provide saliency maps to highlight important input features of a specific eye gaze sequence. However, to date, its localization analysis has been lacking a quantitative approach across entire datasets. In this work, we employ established gaze event detection algorithms for fixations and saccades and quantitatively evaluate the impact of these events by determining their concept influence. Input features that belong to saccades are shown to be substantially more important than features that belong to fixations. By dissecting saccade events into sub-events, we are able to show that gaze samples that are close to the saccadic peak velocity are most influential. We further investigate the effect of event properties like saccadic amplitude or fixational dispersion on the resulting concept influence.

There has been a surge in Explainable-AI (XAI) methods that provide insights into the workings of Deep Neural Network (DNN) models. Integrated Gradients (IG) is a popular XAI algorithm that attributes relevance scores to input features commensurate with their contribution to the model's output. However, it requires multiple forward \& backward passes through the model. Thus, compared to a single forward-pass inference, there is a significant computational overhead to generate the explanation which hinders real-time XAI. This work addresses the aforementioned issue by accelerating IG with a hardware-aware algorithm optimization. We propose a novel non-uniform interpolation scheme to compute the IG attribution scores which replaces the baseline uniform interpolation. Our algorithm significantly reduces the total interpolation steps required without adversely impacting convergence. Experiments on the ImageNet dataset using a pre-trained InceptionV3 model demonstrate \textit{2.6-3.6}$\times$ performance speedup on GPU systems for iso-convergence. This includes the minimal \textit{0.2-3.2}\% latency overhead introduced by the pre-processing stage of computing the non-uniform interpolation step-sizes.

In healthcare, AI systems support clinicians and patients in diagnosis, treatment, and monitoring, but many systems' poor explainability remains challenging for practical application. Overcoming this barrier is the goal of explainable AI (XAI). However, an explanation can be perceived differently and, thus, not solve the black-box problem for everyone. The domain of Human-Centered AI deals with this problem by adapting AI to users. We present a user-centered persona concept to evaluate XAI and use it to investigate end-users preferences for various explanation styles and contents in a mobile health stress monitoring application. The results of our online survey show that users' demographics and personality, as well as the type of explanation, impact explanation preferences, indicating that these are essential features for XAI design. We subsumed the results in three prototypical user personas: power-, casual-, and privacy-oriented users. Our insights bring an interactive, human-centered XAI closer to practical application.

A Human-Centric Assessment Framework for AI

Sascha Saralajew, Ammar Shaker, Zhao Xu +5

With the rise of AI systems in real-world applications comes the need for reliable and trustworthy AI. An essential aspect of this are explainable AI systems. However, there is no agreed standard on how explainable AI systems should be assessed. Inspired by the Turing test, we introduce a human-centric assessment framework where a leading domain expert accepts or rejects the solutions of an AI system and another domain expert. By comparing the acceptance rates of provided solutions, we can assess how the AI system performs compared to the domain expert, and whether the AI system's explanations (if provided) are human-understandable. This setup -- comparable to the Turing test -- can serve as a framework for a wide range of human-centric AI system assessments. We demonstrate this by presenting two instantiations: (1) an assessment that measures the classification accuracy of a system with the option to incorporate label uncertainties; (2) an assessment where the usefulness of provided explanations is determined in a human-centric manner.

Computable Artificial General Intelligence

Michael Timothy Bennett

Artificial general intelligence (AGI) may herald our extinction, according to AI safety research. Yet claims regarding AGI must rely upon mathematical formalisms -- theoretical agents we may analyse or attempt to build. AIXI appears to be the only such formalism supported by proof that its behaviour is optimal, a consequence of its use of compression as a proxy for intelligence. Unfortunately, AIXI is incomputable and claims regarding its behaviour highly subjective. We argue that this is because AIXI formalises cognition as taking place in isolation from the environment in which goals are pursued (Cartesian dualism). We propose an alternative, supported by proof and experiment, which overcomes these problems. Integrating research from cognitive science with AI, we formalise an enactive model of learning and reasoning to address the problem of subjectivity. This allows us to formulate a different proxy for intelligence, called weakness, which addresses the problem of incomputability. We prove optimal behaviour is attained when weakness is maximised. This proof is supplemented by experimental results comparing weakness and description length (the closest analogue to compression possible without reintroducing subjectivity). Weakness outperforms description length, suggesting it is a better proxy. Furthermore we show that, if cognition is enactive, then minimisation of description length is neither necessary nor sufficient to attain optimal performance, undermining the notion that compression is closely related to intelligence. However, there remain open questions regarding the implementation of scale-able AGI. In the short term, these results may be best utilised to improve the performance of existing systems. For example, our results explain why Deepmind's Apperception Engine is able to generalise effectively, and how to replicate that performance by maximising weakness.