Journal of Artificial Intelligence Research

Hierarchy Generation and Exploitation for Enhanced Multi-Class Classification

2026-06-01T09:17:41+00:00

Background: Hierarchical Classification (HC) has long been recognized for improving predictive performance by exploiting relationships between classes. However, most tabular multi-class datasets lack predefined class hierarchies, limiting the broader applicability of hierarchy-aware learning methods.

Objectives: This study introduces HiGEC (Hierarchy Generation and Exploitation for Classification), a unified framework that enables hierarchy-aware learning for standard flat-label tabular classification problems through automatically generated class hierarchies. The objectives are to determine whether data-driven hierarchies can consistently improve performance over Flat Classification (FC), whether hierarchy-aware evaluation metrics improve under enhanced exploitation strategies, and whether probabilistic aggregation mitigates hierarchical error propagation.

Methods: HiGEC systematically integrates Hierarchy Generation (HG) and Hierarchy Exploitation (HE) within a unified framework. Two enhanced HE schemes are proposed: HE+, which mitigates error propagation through probabilistic path aggregation, and HE+F, which combines hierarchical and flat predictions through calibrated convex fusion. A large-scale benchmark involving 100 tabular datasets and ten classifiers—including gradient boosting, ensemble, generative, instancebased, and transformer-based methods—was conducted. Evaluation included both conventional classification metrics and hierarchy-aware metrics, together with statistical significance testing and runtime analysis.

Results: HiGEC consistently improved predictive performance over FC baselines, particularly in F1-score, with statistically significant gains across diverse classifiers. Hierarchy-aware evaluation further demonstrated improvements in structural consistency, with enhanced HE schemes achieving higher hierarchical F-measure (hF) and lower path-based loss than baseline hierarchical methods. Runtime analysis showed that optimized HE+F configurations typically incurred moderate computational overhead (approximately 2–3× slower than FC), while certain lightweight HE+ configurations achieved both higher predictive performance and lower runtime than their FC counterparts.

Conclusions: The empirical findings support the three formal hypotheses of this study: (1) HiGEC improves flat predictive performance over FC, (2) enhanced HE strategies improve hierarchy-aware evaluation metrics, and (3) probabilistic aggregation mitigates hierarchical error propagation. Factor-wise ranking analysis further demonstrated that HE strategies contribute more strongly to performance variation than HG methods under aggregated evaluation. Overall, HiGEC provides a unified and reproducible framework for hierarchy-aware learning on tabular multi-class benchmarks.

Symbol Grounding in Neuro-Symbolic AI: A Gentle Introduction to Reasoning Shortcuts

2026-04-17T07:44:38+00:00

Neuro-symbolic (NeSy) AI aims to develop deep neural networks whose predictions comply with prior knowledge encoding, e.g., safety or structural constraints. As such, it represents one of the most promising avenues for reliable and trustworthy AI. The core idea behind NeSy AI is to combine neural and symbolic steps: neural networks are typically responsible for mapping low-level inputs into high-level symbolic concepts, while symbolic reasoning infers predictions compatible with the extracted concepts and the prior knowledge. Despite their promise, it was recently shown that – whenever the concepts are not supervised directly – NeSy models can be affected by Reasoning Shortcuts (RSs). That is, they can achieve high label accuracy by grounding the concepts incorrectly. RSs can compromise the interpretability of the model’s explanations, performance in out-of-distribution scenarios, and therefore reliability. At the same time, RSs are difficult to detect and prevent unless concept supervision is available, which is typically not the case. However, the literature on RSs is scattered, making it difficult for researchers and practitioners to understand and tackle this challenging problem. This overview addresses this issue by providing a gentle introduction to RSs, discussing their causes and consequences in intuitive terms. It also reviews and elucidates existing theoretical characterizations of this phenomenon. Finally, it details methods for dealing with RSs, including mitigation and awareness strategies, and maps their benefits and limitations. By reformulating advanced material in a digestible form, this overview aims to provide a unifying perspective on RSs to lower the bar to entry for tackling them. Ultimately, we hope this overview contributes to the development of reliable NeSy and trustworthy AI models.

Text Generation: A Systematic Literature Review of Tasks, Evaluation, and Challenges

2026-05-18T08:11:16+00:00

Text generation has become more accessible than ever, and the growing interest in these systems, especially those using large language models, has spurred a surge in related publications. We provide a systematic literature review comprising 257 papers, covering the period from January 2017 to December 2025. This review categorizes text generation contributions into five main tasks: open-ended text generation, summarization, translation, paraphrasing, and question answering. For each task in our taxonomy, we review relevant characteristics and key subtasks. We assess current approaches for evaluating text generation systems, covering model-free, model-based, and human evaluation. Our investigation shows several task-specific challenges (e.g., missing datasets for multi-document summarization, lack of coherence in story generation, and difficulties in complex reasoning for question answering). We further discuss nine challenges common to all tasks and sub-tasks in recent text generation papers: bias, reasoning, hallucinations, misuse, privacy, interpretability, transparency, datasets, and computing. This systematic literature review targets two main audiences: early-career researchers in natural language processing seeking an overview of the field and promising research directions, and senior researchers who need a recent overview of the main tasks, evaluation, challenges, and mitigation strategies.

Symbols and Neurons: A Review of Symbolic XAI in Deep Learning

2026-03-10T00:20:31+00:00

Background: Deep neural networks increasingly power language, vision, and decision systems, yet many deployments require explanations that are faithful, compositional, and governance-ready. Symbolic techniques promise these properties, but the literature mixes post-hoc extraction, knowledge injection, and intrinsically hybrid designs without a unifying view.

Objectives: We provide a systematic review and synthesis of symbolic explainable AI (XAI) for deep learning (January 2017– June 2025), organize the field around a three-part taxonomy—Symbolic Knowledge Extraction (SKE), Symbolic Knowledge Injection (SKI), and Hybrid neurosymbolic architectures—and propose a conceptual framework that clarifies training–inference flows, explanation interfaces, human feedback, and governance touchpoints.

Methods: Beginning from ≈50,000 records, we deduplicated and screened full texts, analyzed 393 PDFs, and included 273 primary studies in the synthesis. We coded each paper for model domain, modality, symbolic formalism, explanation scope and stage, evaluation protocol, and governance alignment. Analyses combine descriptive statistics with stratification by domain and formalism; we qualitatively assess evidence for faithfulness, robustness, data efficiency, and constraint satisfaction.

Results: Research activity accelerates after 2020, with a marked turn toward hybrids. Across the corpus, SKE, SKI, and Hybrid account for approximately 29%, 26%, and 45% of studies, respectively. Rule sets/decision trees remain the dominant explanation artifacts, while logic- and program-based formalisms grow in NLP and planning. SKI most often targets constraint satisfaction and robustness improvements; SKE emphasizes global surrogates and faithfulness auditing; hybrids report gains in sample efficiency and traceable reasoning. However, evaluation practices are heterogeneous, human-subject studies are scarce, and explicit links to policy/risk controls appear in a minority of works.

Conclusions: Our framework unifies how data, priors, and symbolic reasoning interact with neural learners, the explanation interface, human stakeholders, and governance. We distill actionable recommendations: (1) report faithfulness and constraintsatisfaction metrics alongside accuracy; (2) specify symbolic assumptions and training-time injections precisely; (3) include user studies or auditor-centric protocols for high-stakes use; and (4) develop benchmarks that couple tasks with machinereadable knowledge bases. We highlight open problems in scalable formal reasoning with foundation models, verifying generated rationales, and measuring causal faithfulness at scale.

An Empirical Comparison of Cost Functions in Inductive Logic Programming

2026-02-17T10:54:14+00:00

Recent inductive logic programming (ILP) approaches learn optimal hypotheses. An optimal hypothesis minimises a given cost function on the training data. There are many cost functions, such as minimising training error, minimising textual complexity, or minimising the description length of hypotheses. However, selecting an appropriate cost function remains a key question. To address this gap, we extend a constraint-based ILP system to learn optimal hypotheses for seven standard cost functions. We then empirically compare the generalisation error of optimal hypotheses induced under these standard cost functions. Our results on over 20 domains and 1,000 tasks, including game playing, program synthesis, and image reasoning, show that, while no cost function consistently outperforms the others, minimising training error or description length has the best overall performance. Notably, our results indicate that minimising the size of hypotheses does not always reduce generalisation error.

Causal Explanations for Image Classifiers

2026-02-13T08:54:00+00:00

Existing algorithms for explaining the output of image classifiers use different definitions of explanations and a variety of techniques to find them. However, none of the existing tools use a principled approach based on formal definitions of cause and explanation.

In this paper we present a novel black-box approach to computing explanations grounded in the theory of actual causality. We prove relevant theoretical results and present an algorithm for computing approximate explanations based on these definitions. We prove termination of our algorithm and discuss its complexity and the amount of approximation compared to the precise definition.

We implemented the framework in a tool, ReX, and we present experimental results and a comparison with state-of-the-art tools. We demonstrate that ReX is the most efficient black-box tool and produces the smallest explanations, in addition to outperforming other black-box tools on standard quality measures.

VDeH: Voronoi Diagram Encoded Hashing for Effective and Efficient Similarity Search, Applicable to Different Types of Databases

2026-02-20T03:00:30+00:00

The goal of learning to hash (L2H) is to derive data-dependent hash functions from a given data distribution to map data from the input space to a binary coding space. Despite the success of L2H, two observations have cast doubt on the source of its power, i.e., learning. First, a recent study shows that a version of locality-sensitive hashing without learning can achieve comparable accuracy to L2H methods with less time cost. Second, existing L2H methods are constrained to only three types of hash functions: thresholding, hyperspheres, and hyperplanes. In this paper, we identify Voronoi diagrams as a superior alternative for hashing, as they naturally possess three key properties, i.e., full space coverage, entropy maximization, and bit independence, that existing methods must acquire through complex learning. This insight leads us to propose Voronoi Diagram Encoded Hashing (VDeH), a simple and efficient no-learning approach. VDeH constructs hash functions directly from data-driven Voronoi partitions and leverages a simple encoding scheme to generate mutually independent binary bits. The no-learning and data-dependent nature of VDeH makes it an ideal and highly effective plug-and-play component for similarity search, not only in standard vector databases but also for complex data objects such as trajectories and graphs. For complex data objects, VDeH enables a straightforward two-stage process: an appropriate embedding method first maps the data into a vector space, after which VDeH efficiently generates binary hash codes to accelerate retrieval. Comprehensive experiments on multiple large-scale public benchmarks (consisting of databases of vectors, graphs, and trajectories) demonstrate that VDeH significantly outperforms state-of-the-art hashing methods in both retrieval accuracy and computational efficiency, establishing it as a versatile, efficient, and high-performance solution for similarity search across diverse data domains.

Assessing the Minimal Dialectical Quality in Argumentation: A Neuro-Symbolic Approach Integrating Argument Mining, Quality Assessment, and Probabilistic Reasoning

2026-04-02T23:39:24+00:00

This work introduces a new dimension of argumentative quality, termed Minimal Dialectical Quality (MDQ), which requires an argumentation to coherently support its main claims through adequate justifications and, when necessary, explicit rebuttals. MDQ provides a less subjective notion of argumentation quality than many existing approaches, as it relies exclusively on information contained in the text itself and does not require external knowledge or fact-checking. To investigate the feasibility of MDQ as an approximation of human judgments of argumentative quality, we propose the Minimal Dialectical Quality Evaluator (MiDiQE), a neuro-symbolic system for assessing the quality of argumentative essays. MiDiQE operationalizes MDQ by computing the probability that a given text satisfies the new dimension and comparing it against a predefined threshold. The system integrates neural Argumentation Mining, sub-symbolic Single Argument Quality assessment, and symbolic Argumentation Reasoning modules to extract argument structures, evaluate their components, and perform formal probabilistic reasoning over them. Experimental results on argumentative essay datasets show that MDQ aligns well with human quality assessments, supporting its validity as a meaningful quality dimension. They also demonstrate that MiDiQE produces interpretable evaluations of argumentative quality while highlighting both the strengths and limitations of the proposed approach. Potential applications and directions for future work are discussed.

Variety-Seeking Jump Games on Graphs

2026-02-19T02:43:00+00:00

We consider a class of jump games in which agents of different types occupy the nodes of a graph aiming to maximize the variety of types in their neighborhood. In particular, each agent derives a utility equal to the number of types different from its own in its neighborhood. We show that the jump game induced by the strategic behavior of the agents (who aim to maximize their utility) may in general have improving response cycles, but is a potential game under any of the following four conditions: there are only two types of agents; or exactly one empty node; or the graph is of degree at most 2; or the graph is 3-regular and there are two empty nodes. Additionally, we show that on trees, cylinder graphs, and tori, there is always an equilibrium. Finally, we show tight bounds on the price of anarchy with respect to two different measures of diversity: the social welfare (the total utility of the agents) and the number of colorful edges (that connect agents of different types).

Honey, I Shrunk the Hypothesis Space (Through Logical Preprocessing)

2026-01-29T00:07:47+00:00

Inductive logic programming (ILP) is a form of logical machine learning. The goal is to search a hypothesis space for a hypothesis that generalises training examples and background knowledge. We introduce an approach that shrinks the hypothesis space before an ILP system searches it. Our approach uses background knowledge to find rules that cannot be in an optimal hypothesis regardless of the training examples. For instance, our approach discovers relationships such as even numbers cannot be odd and prime numbers greater than 2 are odd. It then removes violating rules from the hypothesis space. We implement our approach using answer set programming and use it to shrink the hypothesis space of a constraint-based ILP system. Our experiments on multiple domains, including visual reasoning and game playing, show that our approach can substantially reduce learning times whilst maintaining predictive accuracies. For instance, given just 10 seconds of preprocessing time, our approach can reduce learning times from over 10 hours to only 2 seconds.