Archimedes at EMNLP 2026

Archimedes members will present five papers at the Conference on Empirical Methods in Natural Language Processing (EMNLP 2026), one of the leading international conferences in natural language processing, taking place in Budapest, Hungary, from 24 to 29 October 2026: two papers in the Findings of EMNLP and three papers at EMNLP workshops.

The accepted papers address cognitive bias in financial text, cross-lingual news clustering, Greek legal retrieval, the reliability of LLM decoding, and multilingual resources for detecting hurtful language.

Findings of EMNLP

• Bias Before the Model: Detecting and Removing Cognitive Bias Triggers in Financial Text
  T. Papadopoulos, C. Venetis, J. Pavlopoulos, I. Androutsopoulos, S. Ananiadou
  arXiv version coming soon
• Cross-lingual Clustering of News Stories using Embeddings-based Community Detection
  N. Nikolaidis, K. Kovacikova, I. Androutsopoulos
  arXiv version coming soon

Workshops

• GreekBarRetrieval: A Benchmark for Greek Statutory Retrieval (NLLP Workshop)
  E. Beta, O. S. Chlapanis, D. Galanis, I. Androutsopoulos
  arXiv:2608.18752
• From Pixels to Tokens: Adaptive LLM Decoding and Reliability Analysis with Otsu Thresholding (UncertaiNLP Workshop)
  E. Partalidou, T. Passali, C. Zerva, G. Tsoumakas, S. Ananiadou
  arXiv version coming soon
• HurtNet: A Multilingual Dictionary of Hurtful Words (OAH Workshop)
  M. A. Roussopoulou, E. D. Palma, A. M. Wizani, P. Osenova, A. Ollagnier, V. Stamou, A. D. Fabio, V. Basile, S. Markantonatou
  arXiv version coming soon

We congratulate all authors and look forward to sharing their work with the NLP community in Budapest.

Learn more about EMNLP 2026: 2026.emnlp.org