Research Papers

Quantifying Explainability of Saliency Methods in Deep Neural Networks with a Synthetic Dataset

Erico Tjoa, Cuntai Guan

Post-hoc analysis is a popular category in eXplainable artificial intelligence (XAI) study. In particular, methods that generate heatmaps have been used to explain the deep neural network (DNN), a black-box model. Heatmaps can be appealing due to the intuitive and visual ways to understand them but assessing their qualities might not be straightforward. Different ways to assess heatmaps' quality have their own merits and shortcomings. This paper introduces a synthetic dataset that can be generated adhoc along with the ground-truth heatmaps for more objective quantitative assessment. Each sample data is an image of a cell with easily recognized features that are distinguished from localization ground-truth mask, hence facilitating a more transparent assessment of different XAI methods. Comparison and recommendations are made, shortcomings are clarified along with suggestions for future research directions to handle the finer details of select post-hoc analysis methods. Furthermore, mabCAM is introduced as the heatmap generation method compatible with our ground-truth heatmaps. The framework is easily generalizable and uses only standard deep learning components.

Gender classification algorithms have important applications in many domains today such as demographic research, law enforcement, as well as human-computer interaction. Recent research showed that algorithms trained on biased benchmark databases could result in algorithmic bias. However, to date, little research has been carried out on gender classification algorithms' bias towards gender minorities subgroups, such as the LGBTQ and the non-binary population, who have distinct characteristics in gender expression. In this paper, we began by conducting surveys on existing benchmark databases for facial recognition and gender classification tasks. We discovered that the current benchmark databases lack representation of gender minority subgroups. We worked on extending the current binary gender classifier to include a non-binary gender class. We did that by assembling two new facial image databases: 1) a racially balanced inclusive database with a subset of LGBTQ population 2) an inclusive-gender database that consists of people with non-binary gender. We worked to increase classification accuracy and mitigate algorithmic biases on our baseline model trained on the augmented benchmark database. Our ensemble model has achieved an overall accuracy score of 90.39%, which is a 38.72% increase from the baseline binary gender classifier trained on Adience. While this is an initial attempt towards mitigating bias in gender classification, more work is needed in modeling gender as a continuum by assembling more inclusive databases.

In today's legal environment, lawsuits and regulatory investigations require companies to embark upon increasingly intensive data-focused engagements to identify, collect and analyze large quantities of data. When documents are staged for review the process can require companies to dedicate an extraordinary level of resources, both with respect to human resources, but also with respect to the use of technology-based techniques to intelligently sift through data. For several years, attorneys have been using a variety of tools to conduct this exercise, and most recently, they are accepting the use of machine learning techniques like text classification to efficiently cull massive volumes of data to identify responsive documents for use in these matters. In recent years, a group of AI and Machine Learning researchers have been actively researching Explainable AI. In an explainable AI system, actions or decisions are human understandable. In typical legal `document review' scenarios, a document can be identified as responsive, as long as one or more of the text snippets in a document are deemed responsive. In these scenarios, if predictive coding can be used to locate these responsive snippets, then attorneys could easily evaluate the model's document classification decision. When deployed with defined and explainable results, predictive coding can drastically enhance the overall quality and speed of the document review process by reducing the time it takes to review documents. The authors of this paper propose the concept of explainable predictive coding and simple explainable predictive coding methods to locate responsive snippets within responsive documents. We also report our preliminary experimental results using the data from an actual legal matter that entailed this type of document review.

COMPL-AI Framework: A Technical Interpretation and LLM Benchmarking Suite for the EU Artificial Intelligence Act

Philipp Guldimann, Alexander Spiridonov, Robin Staab +9

The EU's Artificial Intelligence Act (AI Act) is a significant step towards responsible AI development, but lacks clear technical interpretation, making it difficult to assess models' compliance. This work presents COMPL-AI, a comprehensive framework consisting of (i) the first technical interpretation of the EU AI Act, translating its broad regulatory requirements into measurable technical requirements, with the focus on large language models (LLMs), and (ii) an open-source Act-centered benchmarking suite, based on thorough surveying and implementation of state-of-the-art LLM benchmarks. By evaluating 12 prominent LLMs in the context of COMPL-AI, we reveal shortcomings in existing models and benchmarks, particularly in areas like robustness, safety, diversity, and fairness. This work highlights the need for a shift in focus towards these aspects, encouraging balanced development of LLMs and more comprehensive regulation-aligned benchmarks. Simultaneously, COMPL-AI for the first time demonstrates the possibilities and difficulties of bringing the Act's obligations to a more concrete, technical level. As such, our work can serve as a useful first step towards having actionable recommendations for model providers, and contributes to ongoing efforts of the EU to enable application of the Act, such as the drafting of the GPAI Code of Practice.

Utilizing potent representations of the large vision-language models (VLMs) to accomplish various downstream tasks has attracted increasing attention. Within this research field, soft prompt learning has become a representative approach for efficiently adapting VLMs such as CLIP, to tasks like image classification. However, most existing prompt learning methods learn text tokens that are unexplainable, which cannot satisfy the stringent interpretability requirements of Explainable Artificial Intelligence (XAI) in high-stakes scenarios like healthcare. To address this issue, we propose a novel explainable prompt learning framework that leverages medical knowledge by aligning the semantics of images, learnable prompts, and clinical concept-driven prompts at multiple granularities. Moreover, our framework addresses the lack of valuable concept annotations by eliciting knowledge from large language models and offers both visual and textual explanations for the prompts. Extensive experiments and explainability analyses conducted on various datasets, with and without concept labels, demonstrate that our method simultaneously achieves superior diagnostic performance, flexibility, and interpretability, shedding light on the effectiveness of foundation models in facilitating XAI. The code will be made publically available.

The transition to a sustainable energy supply challenges the operation of electric power systems in manifold ways. Transmission grid loads increase as wind and solar power are often installed far away from the consumers. In extreme cases, system operators must intervene via countertrading or redispatch to ensure grid stability. In this article, we provide a data-driven analysis of congestion in the German transmission grid. We develop an explainable machine learning model to predict the volume of redispatch and countertrade on an hourly basis. The model reveals factors that drive or mitigate grid congestion and quantifies their impact. We show that, as expected, wind power generation is the main driver, but hydropower and cross-border electricity trading also play an essential role. Solar power, on the other hand, has no mitigating effect. Our results suggest that a change to the market design would alleviate congestion.

AI Usage Cards: Responsibly Reporting AI-generated Content

Jan Philip Wahle, Terry Ruas, Saif M. Mohammad +2

Given AI systems like ChatGPT can generate content that is indistinguishable from human-made work, the responsible use of this technology is a growing concern. Although understanding the benefits and harms of using AI systems requires more time, their rapid and indiscriminate adoption in practice is a reality. Currently, we lack a common framework and language to define and report the responsible use of AI for content generation. Prior work proposed guidelines for using AI in specific scenarios (e.g., robotics or medicine) which are not transferable to conducting and reporting scientific research. Our work makes two contributions: First, we propose a three-dimensional model consisting of transparency, integrity, and accountability to define the responsible use of AI. Second, we introduce ``AI Usage Cards'', a standardized way to report the use of AI in scientific research. Our model and cards allow users to reflect on key principles of responsible AI usage. They also help the research community trace, compare, and question various forms of AI usage and support the development of accepted community norms. The proposed framework and reporting system aims to promote the ethical and responsible use of AI in scientific research and provide a standardized approach for reporting AI usage across different research fields. We also provide a free service to easily generate AI Usage Cards for scientific work via a questionnaire and export them in various machine-readable formats for inclusion in different work products at https://ai-cards.org.

Recently, eXplainable AI (XAI) research has focused on counterfactual explanations as post-hoc justifications for AI-system decisions (e.g. a customer refused a loan might be told: If you asked for a loan with a shorter term, it would have been approved). Counterfactuals explain what changes to the input-features of an AI system change the output-decision. However, there is a sub-type of counterfactual, semi-factuals, that have received less attention in AI (though the Cognitive Sciences have studied them extensively). This paper surveys these literatures to summarise historical and recent breakthroughs in this area. It defines key desiderata for semi-factual XAI and reports benchmark tests of historical algorithms (along with a novel, naieve method) to provide a solid basis for future algorithmic developments.

Over the last few years, convolutional neural networks (CNNs) have dominated the field of computer vision thanks to their ability to extract features and their outstanding performance in classification problems, for example in the automatic analysis of X-rays. Unfortunately, these neural networks are considered black-box algorithms, i.e. it is impossible to understand how the algorithm has achieved the final result. To apply these algorithms in different fields and test how the methodology works, we need to use eXplainable AI techniques. Most of the work in the medical field focuses on binary or multiclass classification problems. However, in many real-life situations, such as chest X-rays, radiological signs of different diseases can appear at the same time. This gives rise to what is known as "multilabel classification problems". A disadvantage of these tasks is class imbalance, i.e. different labels do not have the same number of samples. The main contribution of this paper is a Deep Learning methodology for imbalanced, multilabel chest X-ray datasets. It establishes a baseline for the currently underutilised PadChest dataset and a new eXplainable AI technique based on heatmaps. This technique also includes probabilities and inter-model matching. The results of our system are promising, especially considering the number of labels used. Furthermore, the heatmaps match the expected areas, i.e. they mark the areas that an expert would use to make the decision.

Future advances in AI that automate away human labor may have stark implications for labor markets and inequality. This paper proposes a framework to analyze the effects of specific types of AI systems on the labor market, based on how much labor demand they will create versus displace, while taking into account that productivity gains also make society wealthier and thereby contribute to additional labor demand. This analysis enables ethically-minded companies creating or deploying AI systems as well as researchers and policymakers to take into account the effects of their actions on labor markets and inequality, and therefore to steer progress in AI in a direction that advances shared prosperity and an inclusive economic future for all of humanity.

The Clever Hans Effect in Anomaly Detection

Jacob Kauffmann, Lukas Ruff, Grégoire Montavon +1

The 'Clever Hans' effect occurs when the learned model produces correct predictions based on the 'wrong' features. This effect which undermines the generalization capability of an ML model and goes undetected by standard validation techniques has been frequently observed for supervised learning where the training algorithm leverages spurious correlations in the data. The question whether Clever Hans also occurs in unsupervised learning, and in which form, has received so far almost no attention. Therefore, this paper will contribute an explainable AI (XAI) procedure that can highlight the relevant features used by popular anomaly detection models of different type. Our analysis reveals that the Clever Hans effect is widespread in anomaly detection and occurs in many (unexpected) forms. Interestingly, the observed Clever Hans effects are in this case not so much due to the data, but due to the anomaly detection models themselves whose structure makes them unable to detect the truly relevant features, even though vast amounts of data points are available. Overall, our work contributes a warning against an unrestrained use of existing anomaly detection models in practical applications, but it also points at a possible way out of the Clever Hans dilemma, specifically, by allowing multiple anomaly models to mutually cancel their individual structural weaknesses to jointly produce a better and more trustworthy anomaly detector.

This text discusses several popular explanatory methods that go beyond the error measurements and plots traditionally used to assess machine learning models. Some of the explanatory methods are accepted tools of the trade while others are rigorously derived and backed by long-standing theory. The methods, decision tree surrogate models, individual conditional expectation (ICE) plots, local interpretable model-agnostic explanations (LIME), partial dependence plots, and Shapley explanations, vary in terms of scope, fidelity, and suitable application domain. Along with descriptions of these methods, this text presents real-world usage recommendations supported by a use case and public, in-depth software examples for reproducibility.