Efficient rule mining on large modern knowledge graphs (KGs) is a major challenge due to the exponential search space. Current systems -- especially those aiming for exhaustive mining -- remain resource- and time-consuming. In this paper, we propose MiniAMIE, a rule mining approach based on the AMIE algorithm, which restricts AMIE's language bias and estimates key rule metrics using fast approximations. Our experiments on several KGs illustrate the trade-offs of this design and show that MiniAMIE achieves a substantial speed-up while maintaining some good-quality rules.

Local explainability, an important sub-field of eXplainable AI, focuses on describing the decisions of AI models for individual use cases by providing the underlying relationships between a model's inputs and outputs. While the machine learning community has made substantial progress in improving explanation accuracy and completeness, these explanations are rarely evaluated by the final users. In this paper, we evaluate the impact of various explanation and representation techniques on users' comprehension and confidence. Through a user study on two different domains, we assessed three commonly used local explanation techniques—feature-attribution, rule-based, and counterfactual—and explored how their visual representation—graphical or text-based—influences users' comprehension and trust. Our results show that the choice of explanation technique primarily affects user comprehension, whereas the graphical representation impacts user confidence.

Machine learning techniques, such as deep learning and ensemble methods, are widely used in various domains due to their ability to handle complex real-world tasks. However, their black-box nature has raised multiple concerns about the fairness, trustworthiness, and transparency of computer-assisted decision-making. This has led to the emergence of local post-hoc explainability methods, which offer explanations for individual decisions made by black-box algorithms. Among these methods, Kernel SHAP is widely used due to its model-agnostic nature and its well-founded theoretical framework. Despite these strengths, Kernel SHAP suffers from high instability: different executions of the method with the same inputs can lead to significantly different explanations, which diminishes the relevance of the explanations. The contribution of this paper is two-fold. On the one hand, we show that Kernel SHAP's instability is caused by its stochastic neighbor selection procedure, which we adapt to achieve full stability without compromising explanation fidelity. On the other hand, we show that by restricting the neighbors generation to perturbations of size 1 -- which we call the coalitions of Layer 1 -- we obtain a novel feature-attribution method that is fully stable, computationally efficient, and still meaningful.

With the introduction of (large) language models, there has been significant concern about the unintended bias such models may inherit from their training data. A number of studies have shown that such models propagate gender stereotypes, as well as geographical and racial bias, among other biases. While existing works tackle this issue by preprocessing data and debiasing embeddings, the proposed methods require a lot of computational resources and annotation effort while being limited to certain types of biases. To address these issues, we introduce REFINE-LM, a debiasing method that uses reinforcement learning to handle different types of biases without any fine-tuning. By training a simple model on top of the word probability distribution of a LM, our bias agnostic reinforcement learning method enables model debiasing without human annotations or significant computational resources. Experiments conducted on a wide range of models, including several LMs, show that our method (i) significantly reduces stereotypical biases while preserving LMs performance; (ii) is applicable to different types of biases, generalizing across contexts such as gender, ethnicity, religion, and nationality-based biases; and (iii) it is not expensive to train.

As artificial intelligence (AI) systems become increasingly integrated into everyday life, the need for trustworthiness in these systems has emerged as a critical challenge.This tutorial paper addresses the complexity of building trust in AI systems by exploring recent advances in explainable AI (XAI) and related areas that go beyond mere interpretability.After reviewing recent trends in XAI, we discuss how to control AI systems, align them with societal concerns, and address the robustness, reproducibility, and evaluation concerns inherent in these systems.This review highlights the multifaceted nature of the mechanisms for building trust in AI, and we hope it will pave the way for further research in this area.1 https://digital-strategy.

International audience

Academic advising brings numerous benefits to the mission of Higher Education Institutions. One central and challenging duty of advisors is course recommendation for term planning. This task requires both knowledge of the study programs as well as a thorough analysis of the students’ unique circumstances. Limited time and a large student population make this task overwhelming. As a result, an important body of research has sought to expedite term planning via data-oriented decision-support tools. The impact of such tools on students has been extensively studied. However, the advisors’ perspective remains largely unexplored. We contribute to redressing this gap by studying how a grade prediction tool shapes academic advisors’ approach to course recommendation. We found that while the advisors’ usual strategies tend to prevail, their recommendations largely depend on the advisee’s historical performance. That said, advisors also acknowledge the limitations of grades as a measure of academic success.

Recent technological advances rely on accurate decision support systems that can be perceived as black boxes due to their overwhelming complexity. This lack of transparency can lead to technical, ethical, legal, and trust issues. For example, if the control module of a self-driving car failed at detecting a pedestrian, it becomes crucial to know why the system erred. In some other cases, the decision system may reflect unacceptable biases that can generate distrust. The General Data Protection Regulation (GDPR), approved by the European Parliament in 2018, suggests that individuals should be able to obtain explanations of the decisions made from their data by automated processing, and to challenge those decisions. All these reasons have given rise to the domain of interpretable and explainable AI. AIMLAI aims at gathering researchers, experts and professionals, from inside and outside the domain of AI, interested in the topic of interpretable ML and interpretable AI. The workshop encourages interdisciplinary collaborations, with particular emphasis in knowledge management, Infovis, human computer interaction and psychology. It also welcomes applied research for use cases where interpretability matters. AIMLAI envisions to become a discussion venue for the advent of novel interpretable algorithms and explainability modules that mediate the communication between complex ML/AI systems and users.

The increasing interest in transparent and fair AI systems has propelled the research in explainable AI (XAI). One of the main research lines in XAI is post-hoc explainability, the task of explaining the logic of an already deployed black-box model. This is usually achieved by learning an interpretable surrogate function that approximates the black box. Among the existing explanation paradigms, local linear explanations are one of the most popular due to their simplicity and fidelity. Despite their advantages, linear surrogates may not always be the most adapted method to produce reliable, i.e., unambiguous and faithful explanations. Hence, this paper introduces Adapted Post-hoc Explanations (APE), a novel method that characterizes the decision boundary of a black-box classifier and identifies when a linear model constitutes a reliable explanation. Besides, characterizing the black-box frontier allows us to provide complementary counterfactual explanations. Our experimental evaluation shows that APE identifies accurately the situations where linear surrogates are suitable while also providing meaningful counterfactual explanations.

Course selection is a crucial activity for students as it directly impacts their workload and performance. It is also time-consuming, prone to subjectivity, and often carried out based on incomplete information. This task can, nevertheless, be assisted with computational tools, for instance, by predicting performance based on historical data. We investigate the effects of showing grade predictions to students through an interactive visualization tool. A qualitative study suggests that in the presence of predictions, students may focus too much on maximizing their performance, to the detriment of other factors such as the workload. A follow-up quantitative study explored whether these effects are mitigated by changing how predictions are conveyed. Our observations suggest the presence of a framing effect that induces students to put more effort into course selection when faced with more specific predictions. We discuss these and other findings and outline considerations for designing better data-driven course selection tools.

We are interested in understanding the underlying generation process for long sequences of symbolic events. To do so, we propose COSSU, an algorithm to mine small and meaningful sets of sequential rules. The rules are selected using an MDL-inspired criterion that favors compactness and relies on a novel rule-based encoding scheme for sequences. Our evaluation shows that COSSU can successfully retrieve relevant sets of closed sequential rules from a long sequence. Such rules constitute an interpretable model that exhibits competitive accuracy for the tasks of next-element prediction and classification.

Over the past few years, we have witnessed the emergence of large knowledge graphs built by extracting and combining information from multiple sources. This has propelled many advances in query processing over knowledge graphs, however the aspect of providing provenance explanations for query results has so far been mostly neglected. We therefore propose a novel method, SPARQLprov, based on query rewriting, to compute how-provenance polynomials for SPARQL queries over knowledge graphs. Contrary to existing works, SPARQLprov is system-agnostic and can be applied to standard and already deployed SPARQL engines without the need of customized extensions. We rely on spm-semirings to compute polynomial annotations that respect the property of commutation with homomorphisms on monotonic and non-monotonic SPARQL queries without aggregate functions. Our evaluation on real and synthetic data shows that SPARQLprov over standard engines incurs an acceptable runtime overhead w.r.t. the original query, competing with state-of-the-art solutions for how-provenance computation.

The Third Workshop on "Advances in Interpretable Machine Learning and Artificial Intelligence" (AIMLAI) presents contributions in the fields of (i) interpretable ML and AI, i.e., algorithms that are natively interpretable, and (ii) interpretability modules, i.e., explanation layers on top of black-box models, also called post-hoc interpretability. AIMLAI encourages interdisciplinary collaborations with particular emphasis in knowledge management, infovis, human computer interaction and psychology. It also welcomes applied research for use cases where interpretability matters.

Rule-based explanations are a popular method to understand the rationale behind the answers of complex machine learning (ML) classifiers. Recent approaches, such as Anchors, focus on local explanations based on if-then rules that are applicable in the vicinity of a target instance. This has proved effective at producing faithful explanations, yet anchor-based explanations are not free of limitations. These include long overly specific rules as well as explanations of low fidelity. This work presents two simple methods that can mitigate such issues on tabular and textual data. The first approach proposes a careful selection of the discretization method for numerical attributes in tabular datasets. The second one applies the notion of pertinent negatives to explanations on textual data. Our experimental evaluation shows the positive impact of such methods on the quality of anchor-based explanations.

International audience

Current RDF knowledge bases (KBs) are highly incomplete. This incompleteness is a serious problem both for data users and producers. Users do not have guarantees that queries that are run on a KB deliver complete results. Data producers, on the other hand, are blind about the parts of the KB that are incomplete. Yet, completeness information management is poorly supported in the Semantic Web. No RDF storage engine supports reasoning with completeness statements. Moreover, SPARQL cannot express completeness constraints for queries. Motivated by these observations, this paper offers a vision on completeness-aware RDF querying. Our vision includes (1) the sketch of a method to reason about completeness in RDF knowledge bases, (2) two approaches to represent completeness information for SPARQL queries, and (3) an extension for the SPARQL language to express completeness constraints in queries.

Knowledge bases such as Wikidata, DBpedia, or YAGO contain millions of entities and facts. In some knowledge bases, the correctness of these facts has been evaluated. However, much less is known about their completeness, i.e., the proportion of real facts that the knowledge bases cover. In this work, we investigate different signals to identify the areas where a knowledge base is complete. We show that we can combine these signals in a rule mining approach, which allows us to predict where facts may be missing. We also show that completeness predictions can help other applications such as fact prediction.

Open information extraction approaches have led to the creation of large knowledge bases from the Web. The problem with such methods is that their entities and relations are not canonicalized, leading to redundant and ambiguous facts. For example, they may store {Barack Obama, was born, Honolulu and {Obama, place of birth, Honolulu}. In this paper, we present an approach based on machine learning methods that can canonicalize such Open IE triples, by clustering synonymous names and phrases.

Recent advances in information extraction have led to huge knowledge bases (KBs), which capture knowledge in a machine-readable format. Inductive Logic Programming (ILP) can be used to mine logical rules from the KB. These rules can help deduce and add missing knowledge to the KB. While ILP is a mature field, mining logical rules from KBs is different in two aspects: First, current rule mining systems are easily overwhelmed by the amount of data (state-of-the art systems cannot even run on today's KBs). Second, ILP usually requires counterexamples. KBs, however, implement the open world assumption (OWA), meaning that absent data cannot be used as counterexamples. In this paper, we develop a rule mining model that is explicitly tailored to support the OWA scenario. It is inspired by association rule mining and introduces a novel measure for confidence. Our extensive experiments show that our approach outperforms state-of-the-art approaches in terms of precision and coverage. Furthermore, our system, AMIE, mines rules orders of magnitude faster than state-of-the-art approaches.

With RadSpeech, we aim to build the next generation of intelligent, scalable, and user-friendly semantic search interfaces for the medical imaging domain, based on semantic technologies. Ontology-based knowledge representation is used not only for the image contents, but also for the complex natural language understanding and dialogue management process. This demo shows a speech-based annotation system for radiology images and focuses on a new and effective way to annotate medical image regions with a specific medical, structured, diagnosis while using speech and pointing gestures on the go.