Team Vivek Dhayaal at SemEval-2026 Task 13 Subtask B: Multi-Class Authorship Detection

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

Summary

A system developed for SemEval-2026 Task 10 Subtask 2 addresses conspiracy detection, employing a progressive modeling strategy. This approach compares traditional lexical representations, such as Bag-of-Words and TF-IDF features combined with Logistic Regression and Ridge classifiers, against a fine-tuned DistilRoBERTa transformer model for binary classification. All experiments were conducted using only the official task data in a CPU-only environment, without external datasets or data augmentation. The objective was to achieve acceptable performance while minimizing computational resources and model complexity. The DistilRoBERTa transformer model improved the best lexical baseline's performance from 0.67 to 0.75, demonstrating that competitive conspiracy detection is possible with lightweight and reproducible configurations.

Key takeaway

For NLP engineers developing text classification systems, especially in resource-constrained environments, this work demonstrates that you can achieve competitive performance without extensive computational resources or external datasets. By fine-tuning a lightweight transformer like DistilRoBERTa, you can improve classification accuracy significantly over lexical baselines, even on CPU-only setups. Consider this approach to build efficient, reproducible models for tasks like conspiracy detection, optimizing for both performance and operational cost.

Key insights

Competitive conspiracy detection is achievable with lightweight DistilRoBERTa models and minimal resources.

Principles

• Transformer models outperform lexical baselines in conspiracy detection.
• Resource-constrained environments can yield competitive NLP results.
• Reproducible configurations are key for reliable system development.

Method

A progressive modeling strategy compares Bag-of-Words/TF-IDF with Logistic Regression/Ridge classifiers against a fine-tuned DistilRoBERTa transformer for binary classification on official task data.

In practice

• Fine-tune DistilRoBERTa for binary text classification tasks.
• Utilize CPU-only environments for resource-efficient NLP.
• Prioritize official task data to ensure reproducibility.

Topics

• Conspiracy Detection
• DistilRoBERTa
• Transformer Models
• Text Classification
• Resource-Efficient NLP
• SemEval-2026

Best for: Research Scientist, AI Scientist, NLP Engineer

Related on AIssential

• mdok-style at SemEval-2026 Task 10: Finetuning LLMs for Conspiracy Detection — Data augmentation and self-training can effectively finetune LLMs for specialized text classification tasks with limited data.
• How Far Can Classical NLP Go? From Bag-of-Words to Stacking on Spooky Author Identification — Classical NLP, with careful feature engineering and ensemble stacking, excels at stylistic text classification.
• 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.
• From Logistic Regression to GPT-2: Building a Complete Spam Detection & Sentiment Analysis Pipeline — For imbalanced text classification, prioritize F1-score and Precision-Recall AUC over accuracy and ROC-AUC.
• Multi-Stage Training for Abusive Comment Detection in Indic Languages — Multi-stage training with language preprocessing and model ensembles improves abusive comment detection while minimizing false positives.
• Sem-Detect: Semantic Level Detection of AI Generated Peer-Reviews — AI-generated peer reviews converge semantically, while human reviews, even LLM-refined, retain unique ideas.

Open in AIssential →

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