HalCECE: A Framework for Explainable Hallucination Detection through Conceptual Counterfactuals in Image Captioning

2025-03-01

AI Summary

Hallucination—i.e., generation of semantically inconsistent or unsupported content—remains a critical reliability barrier in vision-language (VL) image captioning. Method: This paper introduces the first concept-based counterfactual framework for explainable hallucination detection in VL models. It integrates semantic minimization with hierarchical, knowledge-driven black-box editing to reliably transform hallucinated captions into non-hallucinated ones. We propose the novel “role hallucination” analysis paradigm to uncover semantic interdependencies among visual concepts and systematically apply concept counterfactuals for fine-grained, semantically grounded hallucination attribution. Results: Evaluated across multiple state-of-the-art VL models, our framework significantly enhances the transparency and trustworthiness of hallucination detection. Beyond producing interpretable editing suggestions, it overcomes the limitations of purely numeric evaluation metrics, establishing a new paradigm for trustworthy assessment of VL systems.

Technology Category

• Computer Vision: Large Vision Models
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

In the dynamic landscape of artificial intelligence, the exploration of hallucinations within vision-language (VL) models emerges as a critical frontier. This work delves into the intricacies of hallucinatory phenomena exhibited by widely used image captioners, unraveling interesting patterns. Specifically, we step upon previously introduced techniques of conceptual counterfactual explanations to address VL hallucinations. The deterministic and efficient nature of the employed conceptual counterfactuals backbone is able to suggest semantically minimal edits driven by hierarchical knowledge, so that the transition from a hallucinated caption to a non-hallucinated one is performed in a black-box manner. HalCECE, our proposed hallucination detection framework is highly interpretable, by providing semantically meaningful edits apart from standalone numbers, while the hierarchical decomposition of hallucinated concepts leads to a thorough hallucination analysis. Another novelty tied to the current work is the investigation of role hallucinations, being one of the first works to involve interconnections between visual concepts in hallucination detection. Overall, HalCECE recommends an explainable direction to the crucial field of VL hallucination detection, thus fostering trustworthy evaluation of current and future VL systems.

Problem

• Detects hallucinations in image captioning models
• Uses conceptual counterfactuals for explainable edits
• Analyzes role hallucinations and visual concept interconnections

Innovation

• Conceptual counterfactuals for hallucination detection
• Hierarchical knowledge-driven semantic edits
• Interpretable framework with role hallucination analysis