# Where Detectors Fail: Closing the Tail-Domain Gap with Expert-Guided Mutual Distillation

arXiv: https://arxiv.org/abs/2607.26555 · HTML: https://mlsanity.com/paper/2607.26555/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 13 of its 32 references that are in the map.
- First posted on arXiv 29 Jul 2026.

## Builds on (most cited first, 13 of 13)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [User Preference-aware Fake News Detection](https://mlsanity.com/paper/2104.12259/) (2021)
- [Frequency-Aligned Knowledge Distillation for Lightweight Spatiotemporal Forecasting](https://mlsanity.com/paper/2507.02939/) (2025)
- [Universal Vulnerabilities in Large Language Models: Backdoor Attacks for In-context Learning](https://mlsanity.com/paper/2401.05949/) (2024)
- [Improving Fake News Detection by Using an Entity-enhanced Framework to Fuse Diverse Multimodal Clues](https://mlsanity.com/paper/2108.10509/) (2021)
- [Unlearning Backdoor Attacks for LLMs with Weak-to-Strong Knowledge Distillation](https://mlsanity.com/paper/2410.14425/) (2024)
- [The Authors Matter: Understanding and Mitigating Implicit Bias in Deep Text Classification](https://mlsanity.com/paper/2105.02778/) (2021)
- [COMAE: COMprehensive Attribute Exploration for Zero-shot Hashing](https://mlsanity.com/paper/2402.16424/) (2024)
- [Multi-modal Fake News Detection on Social Media via Multi-grained Information Fusion](https://mlsanity.com/paper/2304.00827/) (2023)
- [Multi-Modal Multi-Behavior Sequential Recommendation with Conditional Diffusion-Based Feature Denoising](https://mlsanity.com/paper/2508.05352/) (2025)
- [Frequency Spectrum is More Effective for Multimodal Representation and Fusion: A Multimodal Spectrum Rumor Detector](https://mlsanity.com/paper/2312.11023/) (2023)
- [C2PO: Diagnosing and Disentangling Bias Shortcuts in LLMs](https://mlsanity.com/paper/2512.23430/) (2025)

## Abstract (arXiv)

Multimodal fake news detectors often generalize poorly across domains because they learn to trust unreliable evidence: domain-specific shortcuts amplified by imbalanced data and semantically inconsistent text-image pairs that make cross-modal evidence unreliable. We propose Expert-Guided Mutual Distillation (EGMD), which learns what evidence to trust across the prediction pipeline. At the input level, input-level calibration encodes pair-level coherence as a shared gain before fusion. At the representation level, an expert-guided teacher aligns domain statistics and encourages domain-specific patterns to concentrate in specialized experts. At the decision level, prototype-anchored domain-specific students use mutual learning and dual-channel distillation to inherit the teacher's feature geometry and calibrated predictions while discouraging local domain priors. We further construct Weibo_Balanced, a domain-balanced benchmark that isolates the effect of imbalance on generalization. Across four datasets in two languages, EGMD achieves state-of-the-art accuracy while reducing domain bias by up to 57.3%.