Research Papers

We propose a new way to explain and to visualize neural network classification through a decomposition-based explainable AI (DXAI). Instead of providing an explanation heatmap, our method yields a decomposition of the image into class-agnostic and class-distinct parts, with respect to the data and chosen classifier. Following a fundamental signal processing paradigm of analysis and synthesis, the original image is the sum of the decomposed parts. We thus obtain a radically different way of explaining classification. The class-agnostic part ideally is composed of all image features which do not posses class information, where the class-distinct part is its complementary. This new visualization can be more helpful and informative in certain scenarios, especially when the attributes are dense, global and additive in nature, for instance, when colors or textures are essential for class distinction. Code is available at https://github.com/dxai2024/dxai.

Understanding when and why to apply any given eXplainable Artificial Intelligence (XAI) technique is not a straightforward task. There is no single approach that is best suited for a given context. This paper aims to address the challenge of selecting the most appropriate explainer given the context in which an explanation is required. For AI explainability to be effective, explanations and how they are presented needs to be oriented towards the stakeholder receiving the explanation. If -- in general -- no single explanation technique surpasses the rest, then reasoning over the available methods is required in order to select one that is context-appropriate. Due to the transparency they afford, we propose employing argumentation techniques to reach an agreement over the most suitable explainers from a given set of possible explainers. In this paper, we propose a modular reasoning system consisting of a given mental model of the relevant stakeholder, a reasoner component that solves the argumentation problem generated by a multi-explainer component, and an AI model that is to be explained suitably to the stakeholder of interest. By formalising supporting premises -- and inferences -- we can map stakeholder characteristics to those of explanation techniques. This allows us to reason over the techniques and prioritise the best one for the given context, while also offering transparency into the selection decision.

Shapley values are extensively used in explainable artificial intelligence (XAI) as a framework to explain predictions made by complex machine learning (ML) models. In this work, we focus on conditional Shapley values for predictive models fitted to tabular data and explain the prediction $f(\boldsymbol{x}^{*})$ for a single observation $\boldsymbol{x}^{*}$ at the time. Numerous Shapley value estimation methods have been proposed and empirically compared on an average basis in the XAI literature. However, less focus has been devoted to analyzing the precision of the Shapley value explanations on an individual basis. We extend our work in Olsen et al. (2023) by demonstrating and discussing that the explanations are systematically less precise for observations on the outer region of the training data distribution for all used estimation methods. This is expected from a statistical point of view, but to the best of our knowledge, it has not been systematically addressed in the Shapley value literature. This is crucial knowledge for Shapley values practitioners, who should be more careful in applying these observations' corresponding Shapley value explanations.

FlexModel: A Framework for Interpretability of Distributed Large Language Models

Matthew Choi, Muhammad Adil Asif, John Willes +1

With the growth of large language models, now incorporating billions of parameters, the hardware prerequisites for their training and deployment have seen a corresponding increase. Although existing tools facilitate model parallelization and distributed training, deeper model interactions, crucial for interpretability and responsible AI techniques, still demand thorough knowledge of distributed computing. This often hinders contributions from researchers with machine learning expertise but limited distributed computing background. Addressing this challenge, we present FlexModel, a software package providing a streamlined interface for engaging with models distributed across multi-GPU and multi-node configurations. The library is compatible with existing model distribution libraries and encapsulates PyTorch models. It exposes user-registerable HookFunctions to facilitate straightforward interaction with distributed model internals, bridging the gap between distributed and single-device model paradigms. Primarily, FlexModel enhances accessibility by democratizing model interactions and promotes more inclusive research in the domain of large-scale neural networks. The package is found at https://github.com/VectorInstitute/flex_model.

Evaluating Trustworthiness of AI-Enabled Decision Support Systems: Validation of the Multisource AI Scorecard Table (MAST)

Pouria Salehi, Yang Ba, Nayoung Kim +10

The Multisource AI Scorecard Table (MAST) is a checklist tool based on analytic tradecraft standards to inform the design and evaluation of trustworthy AI systems. In this study, we evaluate whether MAST is associated with people's trust perceptions in AI-enabled decision support systems (AI-DSSs). Evaluating trust in AI-DSSs poses challenges to researchers and practitioners. These challenges include identifying the components, capabilities, and potential of these systems, many of which are based on the complex deep learning algorithms that drive DSS performance and preclude complete manual inspection. We developed two interactive, AI-DSS test environments using the MAST criteria. One emulated an identity verification task in security screening, and another emulated a text summarization system to aid in an investigative reporting task. Each test environment had one version designed to match low-MAST ratings, and another designed to match high-MAST ratings, with the hypothesis that MAST ratings would be positively related to the trust ratings of these systems. A total of 177 subject matter experts were recruited to interact with and evaluate these systems. Results generally show higher MAST ratings for the high-MAST conditions compared to the low-MAST groups, and that measures of trust perception are highly correlated with the MAST ratings. We conclude that MAST can be a useful tool for designing and evaluating systems that will engender high trust perceptions, including AI-DSS that may be used to support visual screening and text summarization tasks. However, higher MAST ratings may not translate to higher joint performance.

Steering Responsible AI: A Case for Algorithmic Pluralism

Stefaan G. Verhulst

In this paper, I examine questions surrounding AI neutrality through the prism of existing literature and scholarship about mediation and media pluralism. Such traditions, I argue, provide a valuable theoretical framework for how we should approach the (likely) impending era of AI mediation. In particular, I suggest examining further the notion of algorithmic pluralism. Contrasting this notion to the dominant idea of algorithmic transparency, I seek to describe what algorithmic pluralism may be, and present both its opportunities and challenges. Implemented thoughtfully and responsibly, I argue, Algorithmic or AI pluralism has the potential to sustain the diversity, multiplicity, and inclusiveness that are so vital to democracy.

Can governments build AI? In this paper, we describe an ongoing effort to develop ``public AI'' -- publicly accessible AI models funded, provisioned, and governed by governments or other public bodies. Public AI presents both an alternative and a complement to standard regulatory approaches to AI, but it also suggests new technical and policy challenges. We present a roadmap for how the ML research community can help shape this initiative and support its implementation, and how public AI can complement other responsible AI initiatives.

Exploring Practitioner Perspectives On Training Data Attribution Explanations

Elisa Nguyen, Evgenii Kortukov, Jean Y. Song +1

Explainable AI (XAI) aims to provide insight into opaque model reasoning to humans and as such is an interdisciplinary field by nature. In this paper, we interviewed 10 practitioners to understand the possible usability of training data attribution (TDA) explanations and to explore the design space of such an approach. We confirmed that training data quality is often the most important factor for high model performance in practice and model developers mainly rely on their own experience to curate data. End-users expect explanations to enhance their interaction with the model and do not necessarily prioritise but are open to training data as a means of explanation. Within our participants, we found that TDA explanations are not well-known and therefore not used. We urge the community to focus on the utility of TDA techniques from the human-machine collaboration perspective and broaden the TDA evaluation to reflect common use cases in practice.

Explainable AI (XAI) is the study on how humans can be able to understand the cause of a model's prediction. In this work, the problem of interest is Scene Text Recognition (STR) Explainability, using XAI to understand the cause of an STR model's prediction. Recent XAI literatures on STR only provide a simple analysis and do not fully explore other XAI methods. In this study, we specifically work on data explainability frameworks, called attribution-based methods, that explain the important parts of an input data in deep learning models. However, integrating them into STR produces inconsistent and ineffective explanations, because they only explain the model in the global context. To solve this problem, we propose a new method, STRExp, to take into consideration the local explanations, i.e. the individual character prediction explanations. This is then benchmarked across different attribution-based methods on different STR datasets and evaluated across different STR models.

SurroCBM: Concept Bottleneck Surrogate Models for Generative Post-hoc Explanation

Bo Pan, Zhenke Liu, Yifei Zhang +1

Explainable AI seeks to bring light to the decision-making processes of black-box models. Traditional saliency-based methods, while highlighting influential data segments, often lack semantic understanding. Recent advancements, such as Concept Activation Vectors (CAVs) and Concept Bottleneck Models (CBMs), offer concept-based explanations but necessitate human-defined concepts. However, human-annotated concepts are expensive to attain. This paper introduces the Concept Bottleneck Surrogate Models (SurroCBM), a novel framework that aims to explain the black-box models with automatically discovered concepts. SurroCBM identifies shared and unique concepts across various black-box models and employs an explainable surrogate model for post-hoc explanations. An effective training strategy using self-generated data is proposed to enhance explanation quality continuously. Through extensive experiments, we demonstrate the efficacy of SurroCBM in concept discovery and explanation, underscoring its potential in advancing the field of explainable AI.

We propose an approach for early crop classification through identifying important timesteps with eXplainable AI (XAI) methods. Our approach consists of training a baseline crop classification model to carry out layer-wise relevance propagation (LRP) so that the salient time step can be identified. We chose a selected number of such important time indices to create the bounding region of the shortest possible classification timeframe. We identified the period 21st April 2019 to 9th August 2019 as having the best trade-off in terms of accuracy and earliness. This timeframe only suffers a 0.75% loss in accuracy as compared to using the full timeseries. We observed that the LRP-derived important timesteps also highlight small details in input values that differentiates between different classes and

The ability to navigate robots with natural language instructions in an unknown environment is a crucial step for achieving embodied artificial intelligence (AI). With the improving performance of deep neural models proposed in the field of vision-and-language navigation (VLN), it is equally interesting to know what information the models utilize for their decision-making in the navigation tasks. To understand the inner workings of deep neural models, various explanation methods have been developed for promoting explainable AI (XAI). But they are mostly applied to deep neural models for image or text classification tasks and little work has been done in explaining deep neural models for VLN tasks. In this paper, we address these problems by building quantitative benchmarks to evaluate explanation methods for VLN models in terms of faithfulness. We propose a new erasure-based evaluation pipeline to measure the step-wise textual explanation in the sequential decision-making setting. We evaluate several explanation methods for two representative VLN models on two popular VLN datasets and reveal valuable findings through our experiments.