Contextual Stance-Aware Semantic Graph Learning for Fake News Detection

Djamila Benchikh, Mohand Said Allili, Etienne Gael Tajeuna
Proceedings of the The 39th Canadian Conference on Artificial Intelligence, PMLR 318:1005-1011, 2026.

Abstract

The rapid spread of disinformation on social networks threatens public trust and democratic processes. We propose a unified framework for early detection of emerging false narratives by combining context-aware graph modeling with semantic analysis. Our method builds interaction graphs where posts are nodes connected by stance relations (agreement or disagreement). In parallel, a semantic module extracts fine-grained linguistic cues from each post. These signals are fused via a graph neural network that jointly models early diffusion patterns and content semantics to identify deceptive posts at their inception. Experiments on benchmark datasets show that our approach outperforms existing baselines, highlighting the effectiveness of integrating stance-aware graph representations with semantic understanding for scalable disinformation detection.

Cite this Paper

BibTeX
@InProceedings{pmlr-v318-benchikh26a, title = {Contextual Stance-Aware Semantic Graph Learning for Fake News Detection}, author = {Benchikh, Djamila and Allili, Mohand Said and Tajeuna, Etienne Gael}, booktitle = {Proceedings of the The 39th Canadian Conference on Artificial Intelligence}, pages = {1005--1011}, year = {2026}, editor = {Bouzar-Benlabiod, Lydia and Leung, Carson}, volume = {318}, series = {Proceedings of Machine Learning Research}, month = {25--29 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v318/main/assets/benchikh26a/benchikh26a.pdf}, url = {https://proceedings.mlr.press/v318/benchikh26a.html}, abstract = {The rapid spread of disinformation on social networks threatens public trust and democratic processes. We propose a unified framework for early detection of emerging false narratives by combining context-aware graph modeling with semantic analysis. Our method builds interaction graphs where posts are nodes connected by stance relations (agreement or disagreement). In parallel, a semantic module extracts fine-grained linguistic cues from each post. These signals are fused via a graph neural network that jointly models early diffusion patterns and content semantics to identify deceptive posts at their inception. Experiments on benchmark datasets show that our approach outperforms existing baselines, highlighting the effectiveness of integrating stance-aware graph representations with semantic understanding for scalable disinformation detection.} }
Endnote
%0 Conference Paper %T Contextual Stance-Aware Semantic Graph Learning for Fake News Detection %A Djamila Benchikh %A Mohand Said Allili %A Etienne Gael Tajeuna %B Proceedings of the The 39th Canadian Conference on Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Lydia Bouzar-Benlabiod %E Carson Leung %F pmlr-v318-benchikh26a %I PMLR %P 1005--1011 %U https://proceedings.mlr.press/v318/benchikh26a.html %V 318 %X The rapid spread of disinformation on social networks threatens public trust and democratic processes. We propose a unified framework for early detection of emerging false narratives by combining context-aware graph modeling with semantic analysis. Our method builds interaction graphs where posts are nodes connected by stance relations (agreement or disagreement). In parallel, a semantic module extracts fine-grained linguistic cues from each post. These signals are fused via a graph neural network that jointly models early diffusion patterns and content semantics to identify deceptive posts at their inception. Experiments on benchmark datasets show that our approach outperforms existing baselines, highlighting the effectiveness of integrating stance-aware graph representations with semantic understanding for scalable disinformation detection.
APA
Benchikh, D., Allili, M.S. & Tajeuna, E.G.. (2026). Contextual Stance-Aware Semantic Graph Learning for Fake News Detection. Proceedings of the The 39th Canadian Conference on Artificial Intelligence, in Proceedings of Machine Learning Research 318:1005-1011 Available from https://proceedings.mlr.press/v318/benchikh26a.html.

Related Material