Research Papers

The use of Deep Reinforcement Learning (DRL) schemes has increased dramatically since their first introduction in 2015. Though uses in many different applications are being found, they still have a problem with the lack of interpretability. This has bread a lack of understanding and trust in the use of DRL solutions from researchers and the general public. To solve this problem, the field of Explainable Artificial Intelligence (XAI) has emerged. This entails a variety of different methods that look to open the DRL black boxes, ranging from the use of interpretable symbolic Decision Trees (DT) to numerical methods like Shapley Values. This review looks at which methods are being used and for which applications. This is done to identify which models are the best suited to each application or if a method is being underutilised.

Accountability, a requisite for responsible AI, can be facilitated through transparency mechanisms such as audits and explainability. However, prior work suggests that the success of these mechanisms may be limited to Global North contexts; understanding the limitations of current interventions in varied socio-political conditions is crucial to help policymakers facilitate wider accountability. To do so, we examined the mediation of accountability in the existing interactions between vulnerable users and a 'high-risk' AI system in a Global South setting. We report on a qualitative study with 29 financially-stressed users of instant loan platforms in India. We found that users experienced intense feelings of indebtedness for the 'boon' of instant loans, and perceived huge obligations towards loan platforms. Users fulfilled obligations by accepting harsh terms and conditions, over-sharing sensitive data, and paying high fees to unknown and unverified lenders. Users demonstrated a dependence on loan platforms by persisting with such behaviors despite risks of harms such as abuse, recurring debts, discrimination, privacy harms, and self-harm to them. Instead of being enraged with loan platforms, users assumed responsibility for their negative experiences, thus releasing the high-powered loan platforms from accountability obligations. We argue that accountability is shaped by platform-user power relations, and urge caution to policymakers in adopting a purely technical approach to fostering algorithmic accountability. Instead, we call for situated interventions that enhance agency of users, enable meaningful transparency, reconfigure designer-user relations, and prompt a critical reflection in practitioners towards wider accountability. We conclude with implications for responsibly deploying AI in FinTech applications in India and beyond.

The recent success of machine learning (ML) has been fueled by the increasing availability of computing power and large amounts of data in many different applications. However, the trustworthiness of the resulting models can be compromised when such data is maliciously manipulated to mislead the learning process. In this article, we first review poisoning attacks that compromise the training data used to learn ML models, including attacks that aim to reduce the overall performance, manipulate the predictions on specific test samples, and even implant backdoors in the model. We then discuss how to mitigate these attacks using basic security principles, or by deploying ML-oriented defensive mechanisms. We conclude our article by formulating some relevant open challenges which are hindering the development of testing methods and benchmarks suitable for assessing and improving the trustworthiness of ML models against data poisoning attacks

Advances in AI technologies have resulted in superior levels of AI-based model performance. However, this has also led to a greater degree of model complexity, resulting in 'black box' models. In response to the AI black box problem, the field of explainable AI (xAI) has emerged with the aim of providing explanations catered to human understanding, trust, and transparency. Yet, we still have a limited understanding of how xAI addresses the need for explainable AI in the context of healthcare. Our research explores the differing explanation needs amongst stakeholders during the development of an AI-system for classifying COVID-19 patients for the ICU. We demonstrate that there is a constellation of stakeholders who have different explanation needs, not just the 'user'. Further, the findings demonstrate how the need for xAI emerges through concerns associated with specific stakeholder groups i.e., the development team, subject matter experts, decision makers, and the audience. Our findings contribute to the expansion of xAI by highlighting that different stakeholders have different explanation needs. From a practical perspective, the study provides insights on how AI systems can be adjusted to support different stakeholders needs, ensuring better implementation and operation in a healthcare context.

Explainable Artificial Intelligence (XAI)has received a great deal of attention recently. Explainability is being presented as a remedy for the distrust of complex and opaque models. Model agnostic methods such as LIME, SHAP, or Break Down promise instance-level interpretability for any complex machine learning model. But how faithful are these additive explanations? Can we rely on additive explanations for non-additive models? In this paper, we (1) examine the behavior of the most popular instance-level explanations under the presence of interactions, (2) introduce a new method that detects interactions for instance-level explanations, (3) perform a large scale benchmark to see how frequently additive explanations may be misleading.

Generative artificial intelligence (AI) offers numerous opportunities for research and innovation, but its commercialization has raised concerns about the transparency and safety of frontier AI models. Most models lack the necessary components for full understanding, auditing, and reproducibility, and some model producers use restrictive licenses whilst claiming that their models are "open source". To address these concerns, we introduce the Model Openness Framework (MOF), a three-tiered ranked classification system that rates machine learning models based on their completeness and openness, following open science principles. For each MOF class, we specify code, data, and documentation components of the model development lifecycle that must be released and under which open licenses. In addition, the Model Openness Tool (MOT) provides a user-friendly reference implementation to evaluate the openness and completeness of models against the MOF classification system. Together, the MOF and MOT provide timely practical guidance for (i) model producers to enhance the openness and completeness of their publicly-released models, and (ii) model consumers to identify open models and their constituent components that can be permissively used, studied, modified, and redistributed. Through the MOF, we seek to establish completeness and openness as core tenets of responsible AI research and development, and to promote best practices in the burgeoning open AI ecosystem.

Scalable AI Safety via Doubly-Efficient Debate

Jonah Brown-Cohen, Geoffrey Irving, Georgios Piliouras

The emergence of pre-trained AI systems with powerful capabilities across a diverse and ever-increasing set of complex domains has raised a critical challenge for AI safety as tasks can become too complicated for humans to judge directly. Irving et al. [2018] proposed a debate method in this direction with the goal of pitting the power of such AI models against each other until the problem of identifying (mis)-alignment is broken down into a manageable subtask. While the promise of this approach is clear, the original framework was based on the assumption that the honest strategy is able to simulate deterministic AI systems for an exponential number of steps, limiting its applicability. In this paper, we show how to address these challenges by designing a new set of debate protocols where the honest strategy can always succeed using a simulation of a polynomial number of steps, whilst being able to verify the alignment of stochastic AI systems, even when the dishonest strategy is allowed to use exponentially many simulation steps.

The field of eXplainable Artificial Intelligence (XAI) has greatly advanced in recent years, but progress has mainly been made in computer vision and natural language processing. For time series, where the input is often not interpretable, only limited research on XAI is available. In this work, we put forward a virtual inspection layer, that transforms the time series to an interpretable representation and allows to propagate relevance attributions to this representation via local XAI methods like layer-wise relevance propagation (LRP). In this way, we extend the applicability of a family of XAI methods to domains (e.g. speech) where the input is only interpretable after a transformation. Here, we focus on the Fourier transformation which is prominently applied in the interpretation of time series and LRP and refer to our method as DFT-LRP. We demonstrate the usefulness of DFT-LRP in various time series classification settings like audio and electronic health records. We showcase how DFT-LRP reveals differences in the classification strategies of models trained in different domains (e.g., time vs. frequency domain) or helps to discover how models act on spurious correlations in the data.

Multi-dimensional concept discovery (MCD): A unifying framework with completeness guarantees

Johanna Vielhaben, Stefan Blücher, Nils Strodthoff

The completeness axiom renders the explanation of a post-hoc XAI method only locally faithful to the model, i.e. for a single decision. For the trustworthy application of XAI, in particular for high-stake decisions, a more global model understanding is required. Recently, concept-based methods have been proposed, which are however not guaranteed to be bound to the actual model reasoning. To circumvent this problem, we propose Multi-dimensional Concept Discovery (MCD) as an extension of previous approaches that fulfills a completeness relation on the level of concepts. Our method starts from general linear subspaces as concepts and does neither require reinforcing concept interpretability nor re-training of model parts. We propose sparse subspace clustering to discover improved concepts and fully leverage the potential of multi-dimensional subspaces. MCD offers two complementary analysis tools for concepts in input space: (1) concept activation maps, that show where a concept is expressed within a sample, allowing for concept characterization through prototypical samples, and (2) concept relevance heatmaps, that decompose the model decision into concept contributions. Both tools together enable a detailed understanding of the model reasoning, which is guaranteed to relate to the model via a completeness relation. This paves the way towards more trustworthy concept-based XAI. We empirically demonstrate the superiority of MCD against more constrained concept definitions.

Promises and pitfalls of deep neural networks in neuroimaging-based psychiatric research

Fabian Eitel, Marc-André Schulz, Moritz Seiler +2

By promising more accurate diagnostics and individual treatment recommendations, deep neural networks and in particular convolutional neural networks have advanced to a powerful tool in medical imaging. Here, we first give an introduction into methodological key concepts and resulting methodological promises including representation and transfer learning, as well as modelling domain-specific priors. After reviewing recent applications within neuroimaging-based psychiatric research, such as the diagnosis of psychiatric diseases, delineation of disease subtypes, normative modeling, and the development of neuroimaging biomarkers, we discuss current challenges. This includes for example the difficulty of training models on small, heterogeneous and biased data sets, the lack of validity of clinical labels, algorithmic bias, and the influence of confounding variables.

AI systems have seen significant adoption in various domains. At the same time, further adoption in some domains is hindered by inability to fully trust an AI system that it will not harm a human. Besides the concerns for fairness, privacy, transparency, and explainability are key to developing trusts in AI systems. As stated in describing trustworthy AI "Trust comes through understanding. How AI-led decisions are made and what determining factors were included are crucial to understand." The subarea of explaining AI systems has come to be known as XAI. Multiple aspects of an AI system can be explained; these include biases that the data might have, lack of data points in a particular region of the example space, fairness of gathering the data, feature importances, etc. However, besides these, it is critical to have human-centered explanations that are directly related to decision-making similar to how a domain expert makes decisions based on "domain knowledge," that also include well-established, peer-validated explicit guidelines. To understand and validate an AI system's outcomes (such as classification, recommendations, predictions), that lead to developing trust in the AI system, it is necessary to involve explicit domain knowledge that humans understand and use.

Towards Equity and Algorithmic Fairness in Student Grade Prediction

Weijie Jiang, Zachary A. Pardos

Equity of educational outcome and fairness of AI with respect to race have been topics of increasing importance in education. In this work, we address both with empirical evaluations of grade prediction in higher education, an important task to improve curriculum design, plan interventions for academic support, and offer course guidance to students. With fairness as the aim, we trial several strategies for both label and instance balancing to attempt to minimize differences in algorithm performance with respect to race. We find that an adversarial learning approach, combined with grade label balancing, achieved by far the fairest results. With equity of educational outcome as the aim, we trial strategies for boosting predictive performance on historically underserved groups and find success in sampling those groups in inverse proportion to their historic outcomes. With AI-infused technology supports increasingly prevalent on campuses, our methodologies fill a need for frameworks to consider performance trade-offs with respect to sensitive student attributes and allow institutions to instrument their AI resources in ways that are attentive to equity and fairness.