aiqbench analysis
This research addresses critical gaps in multimodal deception detection by introducing auditable reasoning chains and a new multicultural dataset, T4-Deception, to mitigate shortcut learning and improve cross-domain generalization. The proposed architectural modules, including SICS and DMC, enhance robustness under small-data conditions while maintaining state-of-the-art performance across diverse cultural contexts. These contributions provide a more reliable foundation for forensic applications requiring verifiable evidence and consistent behavior beyond narrow laboratory settings.
DecepGPT: Schema-Driven Deception Detection with Multicultural Datasets and Robust Multimodal Learning
arXiv:2603.23916v3 Announce Type: replace-cross Abstract: Multimodal deception detection aims to identify deceptive behavior by analyzing audiovisual cues for forensics and security. In these high-stakes settings, investigators need verifiable evidence connecting audiovisual cues to final decisions, along with reliable generalization across domains and cultural contexts. However, existing benchmarks provide only binary labels without intermediate reasoning cues. Datasets are also small with limited scenario coverage, leading to shortcut learning. We address these issues through three contributions. First, we construct reasoning datasets by augmenting existing benchmarks with structured cue-level descriptions and reasoning chains, enabling model output auditable reports. Second, we release T4-Deception, a multicultural dataset based on the unified ``To Tell The Truth'' television format implemented across four countries. With 1695 samples, it is the largest non-laboratory deception detection dataset. Third, we propose two modules for robust learning under small-data conditions. Stabilized Individuality-Commonality Synergy (SICS) refines multimodal representations by synergizing learnable global priors with sample-adaptive residuals, followed by a polarity-aware adjustment that bi-directionally recalibrates representations. Distilled Modality Consistency (DMC) aligns modality-specific predictions with the fused multimodal predictions via knowledge distillation to prevent unimodal shortcut learning. Experiments on three established benchmarks and our novel dataset demonstrate that our method achieves state-of-the-art performance in both in-domain and cross-domain scenarios, while exhibiting superior transferability across diverse cultural contexts. The datasets and codes will be released.
Abstract:Multimodal deception detection aims to identify deceptive behavior by analyzing audiovisual cues for forensics and security. In these high-stakes settings, investigators need verifiable evidence connecting audiovisual cues to final decisions, along with reliable generalization across domains and cultural contexts. However, existing benchmarks provide only binary labels without intermediate reasoning cues. Datasets are also small with limited scenario coverage, leading to shortcut learning. We address these issues through three contributions. First, we construct reasoning datasets by augmenting existing benchmarks with structured cue-level descriptions and reasoning chains, enabling model output auditable reports. Second, we release T4-Deception, a multicultural dataset based on the unified ``To Tell The Truth'' television format implemented across four countries. With 1695 samples, it is the largest non-laboratory deception detection dataset. Third, we propose two modules for robust learning under small-data conditions. Stabilized Individuality-Commonality Synergy (SICS) refines multimodal representations by synergizing learnable global priors with sample-adaptive residuals, followed by a polarity-aware adjustment that bi-directionally recalibrates representations. Distilled Modality Consistency (DMC) aligns modality-specific predictions with the fused multimodal predictions via knowledge distillation to prevent unimodal shortcut learning. Experiments on three established benchmarks and our novel dataset demonstrate that our method achieves state-of-the-art performance in both in-domain and cross-domain scenarios, while exhibiting superior transferability across diverse cultural contexts. The datasets and codes will be released.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
arXiv:2603.23916 [cs.CV]
(or arXiv:2603.23916v3 [cs.CV] for this version)
From: Jiajian Huang Mr. [view email] [v1] Wed, 25 Mar 2026 04:06:36 UTC (2,638 KB) [v2] Sat, 11 Apr 2026 11:12:51 UTC (2,866 KB) [v3] Sun, 7 Jun 2026 07:53:12 UTC (2,865 KB)
Full text under the publisher's open licence · 264 words
Related coverage
AI-powered metadata correction and harmonization
AWS · Aug 24, 2026
Model-agnostic PII detection with LLMs
AWS · Sep 10, 2026
From weeks to minutes: How Formula 1® uses agentic AI on AWS to accelerate data operations
AWS · Aug 3, 2026
Building trust in AI agents: A production-ready evaluation framework with Amazon Bedrock AgentCore
AWS · Oct 16, 2026