Explainable Artificial Intelligence (XAI)

Derniers papiers

• Paraconsistent-Lib: an intuitive PAL2v algorithm Python Library

  Arnaldo de Carvalho Junior, Diego Oliveira da Cruz, Bruno da Silva Alves, Fernando da Silva Paulo Junior, Jo\~ao Inacio da Silva Filho · 27 novembre 2025

  This paper introduces Paraconsistent-Lib, an open-source, easy-to-use Python library for building PAL2v algorithms in reasoning and decision-making systems. Paraconsistent-Lib is designed as a general-purpose library of PAL2v standard calculations, presenting three types of results: paraconsistent a…

• The Directed Prediction Change - Efficient and Trustworthy Fidelity Assessment for Local Feature Attribution Methods

  Kevin Iselborn, David Dembinsky, Adriano Lucieri, Andreas Dengel · 27 novembre 2025

  The utility of an explanation method critically depends on its fidelity to the underlying machine learning model. Especially in high-stakes medical settings, clinicians and regulators require explanations that faithfully reflect the model's decision process. Existing fidelity metrics such as Infidel…

• Mechanistic Interpretability for Transformer-based Time Series Classification

  Mat\=iss Kaln\=are, Sofoklis Kitharidis, Thomas B\"ack, Niki van Stein · 27 novembre 2025

  Transformer-based models have become state-of-the-art tools in various machine learning tasks, including time series classification, yet their complexity makes understanding their internal decision-making challenging. Existing explainability methods often focus on input-output attributions, leaving …

• Guaranteed Optimal Compositional Explanations for Neurons

  Biagio La Rosa, Leilani H. Gilpin · 27 novembre 2025

  While neurons are the basic units of deep neural networks, it is still unclear what they learn and if their knowledge is aligned with that of humans. Compositional explanations aim to answer this question by describing the spatial alignment between neuron activations and concepts through logical rul…

• Open Vocabulary Compositional Explanations for Neuron Alignment

  Biagio La Rosa, Leilani H. Gilpin · 27 novembre 2025

  Neurons are the fundamental building blocks of deep neural networks, and their interconnections allow AI to achieve unprecedented results. Motivated by the goal of understanding how neurons encode information, compositional explanations leverage logical relationships between concepts to express the …

• CHiQPM: Calibrated Hierarchical Interpretable Image Classification

  Thomas Norrenbrock, Timo Kaiser, Sovan Biswas, Neslihan Kose, Ramesh Manuvinakurike, Bodo Rosenhahn · 27 novembre 2025

  Globally interpretable models are a promising approach for trustworthy AI in safety-critical domains. Alongside global explanations, detailed local explanations are a crucial complement to effectively support human experts during inference. This work proposes the Calibrated Hierarchical QPM (CHiQPM)…

• IVY-FAKE: A Unified Explainable Framework and Benchmark for Image and Video AIGC Detection

  Changjiang Jiang, Wenhui Dong, Zhonghao Zhang, Chenyang Si, Fengchang Yu, Wei Peng, Xinbin Yuan, Yifei Bi, Ming Zhao, Zian Zhou, Caifeng Shan · 26 novembre 2025

  The rapid development of Artificial Intelligence Generated Content (AIGC) techniques has enabled the creation of high-quality synthetic content, but it also raises significant security concerns. Current detection methods face two major limitations: (1) the lack of multidimensional explainable datase…

• Many Ways to be Right: Rashomon Sets for Concept-Based Neural Networks

  Shihan Feng, Cheng Zhang, Michael Xi, Ethan Hsu, Lesia Semenova, Chudi Zhong · 26 novembre 2025

  Modern neural networks rarely have a single way to be right. For many tasks, multiple models can achieve identical performance while relying on different features or reasoning patterns, a property known as the Rashomon Effect. However, uncovering this diversity in deep architectures is challenging a…

• CoT Red-Handed: Stress Testing Chain-of-Thought Monitoring

  Benjamin Arnav, Pablo Bernabeu-P\'erez, Nathan Helm-Burger, Tim Kostolansky, Hannes Whittingham, Mary Phuong · 26 novembre 2025

  As AI models are deployed with increasing autonomy, it is important to ensure they do not take harmful actions unnoticed. As a potential mitigation, we investigate Chain-of-Thought (CoT) monitoring, wherein a weaker trusted monitor model continuously oversees the intermediate reasoning steps of a mo…

• Actionable and diverse counterfactual explanations incorporating domain knowledge and causal constraints

  Szymon Bobek, {\L}ukasz Ba{\l}ec, Grzegorz J. Nalepa · 26 novembre 2025

  Counterfactual explanations enhance the actionable interpretability of machine learning models by identifying the minimal changes required to achieve a desired outcome of the model. However, existing methods often ignore the complex dependencies in real-world datasets, leading to unrealistic or impr…

• Deep Hidden Cognition Facilitates Reliable Chain-of-Thought Reasoning

  Zijun Chen, Wenbo Hu, Richang Hong · 26 novembre 2025

  Chain of Thought (CoT) reasoning has demonstrated remarkable deep reasoning capabilities in both large language models (LLMs) and multimodal large language models (MLLMs). However, its reliability is often undermined by the accumulation of errors in intermediate steps. This paper introduces an novel…

• Counterfactual Simulatability of LLM Explanations for Generation Tasks

  Marvin Limpijankit, Yanda Chen, Melanie Subbiah, Nicholas Deas, Kathleen McKeown · 26 novembre 2025

  LLMs can be unpredictable, as even slight alterations to the prompt can cause the output to change in unexpected ways. Thus, the ability of models to accurately explain their behavior is critical, especially in high-stakes settings. One approach for evaluating explanations is counterfactual simulata…

• Beyond Components: Singular Vector-Based Interpretability of Transformer Circuits

  Areeb Ahmad, Abhinav Joshi, Ashutosh Modi · 26 novembre 2025

  Transformer-based language models exhibit complex and distributed behavior, yet their internal computations remain poorly understood. Existing mechanistic interpretability methods typically treat attention heads and multilayer perceptron layers (MLPs) (the building blocks of a transformer architectu…

• Progressive Localisation in Localist LLMs

  Joachim Diederich · 25 novembre 2025

  This paper demonstrates that progressive localization, the gradual increase of attention locality from early distributed layers to late localized layers, represents the optimal architecture for creating interpretable large language models while preserving performance. Through systematic experimentat…

• XAI-on-RAN: Explainable, AI-native, and GPU-Accelerated RAN Towards 6G

  Osman Tugay Basaran, Falko Dressler · 25 novembre 2025

  Artificial intelligence (AI)-native radio access networks (RANs) will serve vertical industries with stringent requirements: smart grids, autonomous vehicles, remote healthcare, industrial automation, etc. To achieve these requirements, modern 5G/6G design increasingly leverage AI for network optimi…

• RAVEN++: Pinpointing Fine-Grained Violations in Advertisement Videos with Active Reinforcement Reasoning

  Deyi Ji, Yuekui Yang, Liqun Liu, Peng Shu, Haiyang Wu, Shaogang Tang, Xudong Chen, Shaoping Ma, Tianrun Chen, Lanyun Zhu · 25 novembre 2025

  Advertising (Ad) is a cornerstone of the digital economy, yet the moderation of video advertisements remains a significant challenge due to their complexity and the need for precise violation localization. While recent advancements, such as the RAVEN model, have improved coarse-grained violation det…

• Fidelity-Aware Recommendation Explanations via Stochastic Path Integration

  Oren Barkan, Yahlly Schein, Yehonatan Elisha, Veronika Bogina, Mikhail Baklanov, Noam Koenigstein · 25 novembre 2025

  Explanation fidelity, which measures how accurately an explanation reflects a model's true reasoning, remains critically underexplored in recommender systems. We introduce SPINRec (Stochastic Path Integration for Neural Recommender Explanations), a model-agnostic approach that adapts path-integratio…

• Findings of the BlackboxNLP 2025 Shared Task: Localizing Circuits and Causal Variables in Language Models

  Dana Arad, Yonatan Belinkov, Hanjie Chen, Najoung Kim, Hosein Mohebbi, Aaron Mueller, Gabriele Sarti, Martin Tutek · 25 novembre 2025

  Mechanistic interpretability (MI) seeks to uncover how language models (LMs) implement specific behaviors, yet measuring progress in MI remains challenging. The recently released Mechanistic Interpretability Benchmark (MIB; Mueller et al., 2025) provides a standardized framework for evaluating circu…

• Empathetic Cascading Networks: A Multi-Stage Prompting Technique for Reducing Social Biases in Large Language Models

  Wangjiaxuan Xin · 25 novembre 2025

  This report presents the Empathetic Cascading Networks (ECN) framework, a multi-stage prompting method designed to enhance the empathetic and inclusive capabilities of large language models. ECN employs four stages: Perspective Adoption, Emotional Resonance, Reflective Understanding, and Integrative…

• Causally Reliable Concept Bottleneck Models

  Giovanni De Felice, Arianna Casanova Flores, Francesco De Santis, Silvia Santini, Johannes Schneider, Pietro Barbiero, Alberto Termine · 25 novembre 2025

  Concept-based models are an emerging paradigm in deep learning that constrains the inference process to operate through human-interpretable variables, facilitating explainability and human interaction. However, these architectures, on par with popular opaque neural models, fail to account for the tr…

• Reversing the Lens: Using Explainable AI to Understand Human Expertise

  Roussel Rahman, Aashwin Ananda Mishra, Wan-Lin Hu · 25 novembre 2025

  Both humans and machine learning models learn from experience, particularly in safety- and reliability-critical domains. While psychology seeks to understand human cognition, the field of Explainable AI (XAI) develops methods to interpret machine learning models. This study bridges these domains by …

• Multi-Value Alignment for LLMs via Value Decorrelation and Extrapolation

  Hefei Xu, Le Wu, Chen Cheng, Hao Liu · 25 novembre 2025

  With the rapid advancement of large language models (LLMs), aligning them with human values for safety and ethics has become a critical challenge. This problem is especially challenging when multiple, potentially conflicting human values must be considered and balanced. Although several variants of …

• AI Debaters are More Persuasive when Arguing in Alignment with Their Own Beliefs

  Mar\'ia Victoria Carro, Denise Alejandra Mester, Facundo Nieto, Oscar Agust\'in Stanchi, Guido Ernesto Bergman, Mario Alejandro Leiva, Eitan Sprejer, Luca Nicol\'as Forziati Gangi, Francisca Gauna Selasco, Juan Gustavo Corval\'an, Gerardo I. Simari, Mar\'ia Vanina Martinez · 25 novembre 2025

  The core premise of AI debate as a scalable oversight technique is that it is harder to lie convincingly than to refute a lie, enabling the judge to identify the correct position. Yet, existing debate experiments have relied on datasets with ground truth, where lying is reduced to defending an incor…

• BlockCert: Certified Blockwise Extraction of Transformer Mechanisms

  Sandro Andric · 25 novembre 2025

  Mechanistic interpretability aspires to reverse-engineer neural networks into explicit algorithms, while model editing seeks to modify specific behaviours without retraining. Both areas are typically evaluated with informal evidence and ad-hoc experiments, with few explicit guarantees about how far …

• Unboxing the Black Box: Mechanistic Interpretability for Algorithmic Understanding of Neural Networks

  Bianka Kowalska, Halina Kwa\'snicka · 25 novembre 2025

  The black box nature of deep neural networks poses a significant challenge for the deployment of transparent and trustworthy artificial intelligence (AI) systems. With the growing presence of AI in society, it becomes increasingly important to develop methods that can explain and interpret the decis…