Two Outliers at BEA 2025 Shared Task: Tutor Identity Classification using DiReC, a Two-Stage Disentangled Contrastive Representation

Eduardus Tjitrahardja, Ikhlasul Hanif


Abstract
This paper presents DiReC (Disentangled Contrastive Representation), a novel two-stage framework designed to address the BEA 2025 Shared Task 5: Tutor Identity Classification. The task involves distinguishing between responses generated by nine different tutors, including both human educators and large language models (LLMs). DiReC leverages a disentangled representation learning approach, separating semantic content and stylistic features to improve tutor identification accuracy. In Stage 1, the model learns discriminative content representations using cross-entropy loss. In Stage 2, it applies supervised contrastive learning on style embeddings and introduces a disentanglement loss to enforce orthogonality between style and content spaces. Evaluated on the validation set, DiReC achieves strong performance, with a macro-F1 score of 0.9101 when combined with a CatBoost classifier and refined using the Hungarian algorithm. The system ranks third overall in the shared task with a macro-F1 score of 0.9172, demonstrating the effectiveness of disentangled representation learning for tutor identity classification.
Anthology ID:
2025.bea-1.97
Volume:
Proceedings of the 20th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2025)
Month:
July
Year:
2025
Address:
Vienna, Austria
Editors:
Ekaterina Kochmar, Bashar Alhafni, Marie Bexte, Jill Burstein, Andrea Horbach, Ronja Laarmann-Quante, Anaïs Tack, Victoria Yaneva, Zheng Yuan
Venues:
BEA | WS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1212–1223
Language:
URL:
https://preview.aclanthology.org/acl25-workshop-ingestion/2025.bea-1.97/
DOI:
Bibkey:
Cite (ACL):
Eduardus Tjitrahardja and Ikhlasul Hanif. 2025. Two Outliers at BEA 2025 Shared Task: Tutor Identity Classification using DiReC, a Two-Stage Disentangled Contrastive Representation. In Proceedings of the 20th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2025), pages 1212–1223, Vienna, Austria. Association for Computational Linguistics.
Cite (Informal):
Two Outliers at BEA 2025 Shared Task: Tutor Identity Classification using DiReC, a Two-Stage Disentangled Contrastive Representation (Tjitrahardja & Hanif, BEA 2025)
Copy Citation:
PDF:
https://preview.aclanthology.org/acl25-workshop-ingestion/2025.bea-1.97.pdf