# STEM: Unsupervised STructural EMbedding for Stance Detection

arXiv: https://arxiv.org/abs/2112.00712 · HTML: https://mlsanity.com/paper/2112.00712/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 52 references that are in the map.
- First posted on arXiv 1 Dec 2021.

## Builds on (most cited first, 12 of 12)

- [Zero-Shot Stance Detection: A Dataset and Model using Generalized Topic Representations](https://mlsanity.com/paper/2010.03640/) (2020)
- [SemEval-2017 Task 8: RumourEval: Determining rumour veracity and support for rumours](https://mlsanity.com/paper/1704.05972/) (2017)
- [Will-They-Won't-They: A Very Large Dataset for Stance Detection on Twitter](https://mlsanity.com/paper/2005.00388/) (2020)
- [A Retrospective Analysis of the Fake News Challenge Stance Detection Task](https://mlsanity.com/paper/1806.05180/) (2018)
- [Detecting Stance in Media on Global Warming](https://mlsanity.com/paper/2010.15149/) (2020)
- [Modeling Conversation Structure and Temporal Dynamics for Jointly Predicting Rumor Stance and Veracity](https://mlsanity.com/paper/1909.08211/) (2019)
- [Stance Classification in Rumours as a Sequential Task Exploiting the Tree Structure of Social Media Conversations](https://mlsanity.com/paper/1609.09028/) (2016)
- [Adversarial Domain Adaptation for Stance Detection](https://mlsanity.com/paper/1902.02401/) (2019)
- [ConStance: Modeling Annotation Contexts to Improve Stance Classification](https://mlsanity.com/paper/1708.06309/) (2017)
- [Discourse Parsing of Contentious, Non-Convergent Online Discussions](https://mlsanity.com/paper/2012.04585/) (2020)
- [Improving Agreement and Disagreement Identification in Online Discussions with A Socially-Tuned Sentiment Lexicon](https://mlsanity.com/paper/1606.05706/) (2016)
- [Agree to Disagree: Improving Disagreement Detection with Dual GRUs](https://mlsanity.com/paper/1708.05582/) (2017)

## Built on by (most cited first, 9 of 9)

- [Prompting and Fine-Tuning Open-Sourced Large Language Models for Stance Classification](https://mlsanity.com/paper/2309.13734/) (2023)
- [Large Language Models Meet Stance Detection: A Survey of Tasks, Methods, Applications, Challenges and Future Directions](https://mlsanity.com/paper/2505.08464/) (2025)
- [Reinforcement Tuning for Detecting Stances and Debunking Rumors Jointly with Large Language Models](https://mlsanity.com/paper/2406.02143/) (2024)
- [DoubleH: Twitter User Stance Detection via Bipartite Graph Neural Networks](https://mlsanity.com/paper/2301.08774/) (2023)
- [MSME: A Multi-Stage Multi-Expert Framework for Zero-Shot Stance Detection](https://mlsanity.com/paper/2512.04492/) (2025)
- [The Branch Not Taken: Predicting Branching in Online Conversations](https://mlsanity.com/paper/2404.13613/) (2024)
- [When Misinformation Speaks and Converses: Rethinking Fact-Checking in Audio Platforms](https://mlsanity.com/paper/2604.16767/) (2026)
- [Contextualized Prompting For Stance Detection On Social Media](https://mlsanity.com/paper/2606.06022/) (2026)
- [ValueGraph: Value-Signal Guided Graph Pre-training for Contextualized User Representation](https://mlsanity.com/paper/2609.00057/) (2026)

## Abstract (arXiv)

Stance detection is an important task, supporting many downstream tasks such as discourse parsing and modeling the propagation of fake news, rumors, and science denial. In this paper, we propose a novel framework for stance detection. Our framework is unsupervised and domain-independent. Given a claim and a multi-participant discussion - we construct the interaction network from which we derive topological embedding for each speaker. These speaker embedding enjoy the following property: speakers with the same stance tend to be represented by similar vectors, while antipodal vectors represent speakers with opposing stances. These embedding are then used to divide the speakers into stance-partitions. We evaluate our method on three different datasets from different platforms. Our method outperforms or is comparable with supervised models while providing confidence levels for its output. Furthermore, we demonstrate how the structural embedding relate to the valence expressed by the speakers. Finally, we discuss some limitations inherent to the framework.