TTLab at SemEval-2026 Task 10: Transformer-based Approaches for Psycholinguistic Conspiracy Detection in Social Media Discourse

2026-06-25 · Source: Paper Index on ACL Anthology · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Social Sciences & Behavioral Studies · Depth: Expert, short

Summary

TTLab participated in SemEval-2026 Task 10 (PsyCoMark), focusing on automatically detecting conspiracy narratives within Reddit conversations. The team evaluated four distinct transformer-based approaches: using raw text, incorporating structured psycholinguistic markers, a combined representation of both, and a stacking ensemble model. Their findings, presented at the 20th International Workshop on Semantic Evaluation in July 2026, demonstrated that models leveraging psycholinguistic markers significantly outperformed those relying solely on raw text. Furthermore, the implementation of a stacking ensemble approach enhanced the overall robustness of the detection system. This work highlights the critical role of integrating structured psychological cues to develop scalable and reliable conspiracy detection systems.

Key takeaway

For NLP Engineers developing social media content moderation systems, consider integrating psycholinguistic markers into your transformer models for improved conspiracy detection. Your systems will likely achieve higher accuracy and robustness by combining raw text analysis with structured psychological cues, especially when employing ensemble methods. This approach offers a scalable path to more reliable identification of harmful narratives on platforms like Reddit.

Key insights

Incorporating psycholinguistic markers significantly enhances transformer-based conspiracy detection in social media.

Principles

• Psycholinguistic cues improve detection.
• Ensembling boosts model robustness.
• Structured data aids text analysis.

Method

TTLab evaluated four transformer-based approaches: raw text, structured psycholinguistic markers, a combined representation, and a stacking ensemble for conspiracy detection in Reddit conversations.

In practice

• Integrate psycholinguistic features.
• Apply ensemble learning for stability.
• Analyze Reddit for narrative patterns.

Topics

• Conspiracy Detection
• Transformer Models
• Psycholinguistic Markers
• Social Media Analysis
• Reddit Discourse
• Ensemble Learning

Best for: AI Scientist, NLP Engineer, Research Scientist

Related on AIssential

• Beyond Surface Forms: A Comprehensive, Mechanism-Oriented Taxonomy of Indirect Linguistic Encoding for LLM-Based Coded Language Detection — A mechanism-oriented taxonomy of indirect linguistic encoding significantly enhances LLM-based coded language detection.
• Understanding Sarcasm in Generative AI Models: How ChatGPT Interprets Human Intent — Sarcasm detection in AI requires understanding non-literal meaning through contextual learning, moving beyond simple keyword analysis.
• Robust Dual-Signal Fusion: Hybrid Neuro-Symbolic Gating with Compressed Chain-of-Thought Refinement for Irony Detection in Social Media Texts — Hybrid neuro-symbolic fusion with compressed Chain-of-Thought effectively overcomes LLM literalism for zero-shot irony detection.
• Fake News Detection Using Natural Language Processing! — Combining linguistic, syntactic, and semantic features with machine learning achieves high accuracy in fake news detection.
• A Multitask Transformer for Offensive Language Detection and Target Identification in HateBR — Multitask learning improves offensive language detection consistency but may reduce fine-grained target attribution.
• An End-to-End Hybrid Framework for Rumour Detection in Low-Resources Algerian Dialect — A hybrid framework combining transformer embeddings with classical classifiers enables effective rumour detection in low-resource Algerian dialect.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Paper Index on ACL Anthology.