Multi-Task Network Guided Multimodal Fusion for Fake News Detection

Jinke Ma, Liyuan Zhang, Yong Liu, Wei Zhang
Proceedings of the 16th Asian Conference on Machine Learning, PMLR 260:813-828, 2025.

Abstract

Fake news detection has become a hot research topic in the multimodal domain. Existing multimodal fake news detection research utilizes a series of feature fusion networks to gather useful information from different modalities of news posts. However, how to form effective cross-modal features? And how cross-modal correlations impact decision-making? These remain open questions. This paper introduces MMFND, a multi-task guided multimodal fusion framework for fake news detection , which introduces multi-task modules for feature refinement and fusion. Pairwise CLIP encoders are used to extract modality-aligned deep representations, enabling accurate measurement of cross-modal correlations. Enhancing feature fusion by weighting multimodal features with normalised cross-modal correlations. Extensive experiments on typical fake news datasets demonstrate that MMFND outperforms state-of-the-art approaches.

Cite this Paper

BibTeX
@InProceedings{pmlr-v260-ma25a, title = {Multi-Task Network Guided Multimodal Fusion for Fake News Detection}, author = {Ma, Jinke and Zhang, Liyuan and Liu, Yong and Zhang, Wei}, booktitle = {Proceedings of the 16th Asian Conference on Machine Learning}, pages = {813--828}, year = {2025}, editor = {Nguyen, Vu and Lin, Hsuan-Tien}, volume = {260}, series = {Proceedings of Machine Learning Research}, month = {05--08 Dec}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v260/main/assets/ma25a/ma25a.pdf}, url = {https://proceedings.mlr.press/v260/ma25a.html}, abstract = {Fake news detection has become a hot research topic in the multimodal domain. Existing multimodal fake news detection research utilizes a series of feature fusion networks to gather useful information from different modalities of news posts. However, how to form effective cross-modal features? And how cross-modal correlations impact decision-making? These remain open questions. This paper introduces MMFND, a multi-task guided multimodal fusion framework for fake news detection , which introduces multi-task modules for feature refinement and fusion. Pairwise CLIP encoders are used to extract modality-aligned deep representations, enabling accurate measurement of cross-modal correlations. Enhancing feature fusion by weighting multimodal features with normalised cross-modal correlations. Extensive experiments on typical fake news datasets demonstrate that MMFND outperforms state-of-the-art approaches.} }
Endnote
%0 Conference Paper %T Multi-Task Network Guided Multimodal Fusion for Fake News Detection %A Jinke Ma %A Liyuan Zhang %A Yong Liu %A Wei Zhang %B Proceedings of the 16th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Vu Nguyen %E Hsuan-Tien Lin %F pmlr-v260-ma25a %I PMLR %P 813--828 %U https://proceedings.mlr.press/v260/ma25a.html %V 260 %X Fake news detection has become a hot research topic in the multimodal domain. Existing multimodal fake news detection research utilizes a series of feature fusion networks to gather useful information from different modalities of news posts. However, how to form effective cross-modal features? And how cross-modal correlations impact decision-making? These remain open questions. This paper introduces MMFND, a multi-task guided multimodal fusion framework for fake news detection , which introduces multi-task modules for feature refinement and fusion. Pairwise CLIP encoders are used to extract modality-aligned deep representations, enabling accurate measurement of cross-modal correlations. Enhancing feature fusion by weighting multimodal features with normalised cross-modal correlations. Extensive experiments on typical fake news datasets demonstrate that MMFND outperforms state-of-the-art approaches.
APA
Ma, J., Zhang, L., Liu, Y. & Zhang, W.. (2025). Multi-Task Network Guided Multimodal Fusion for Fake News Detection. Proceedings of the 16th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 260:813-828 Available from https://proceedings.mlr.press/v260/ma25a.html.

Related Material