Promoting generalized cross-lingual question answering in few-resource scenarios via self-knowledge distillation

We address the challenge of Generalized Cross-Lingual Transfer (G-XLT) in extractive Question Answering, where question and context languages differ, a problem particularly difficult for low-resource languages. Working with only a thousand parallel QA samples, we combine cross-lingual sampling with...

Full description

Bibliographic Details
Authors: Carrino, Casimiro Pio, Escolano Peinado, Carlos|||0000-0001-6657-673X, Gasco Sánchez, Luis, Rodríguez Fonollosa, José Adrián|||0000-0001-9513-7939
Format: article
Publication Date:2025
Country:España
Institution:Universitat Politècnica de Catalunya (UPC)
Repository:UPCommons. Portal del coneixement obert de la UPC
Language:English
OAI Identifier:oai:upcommons.upc.edu:2117/449061
Online Access:https://hdl.handle.net/2117/449061
https://dx.doi.org/10.26342/2025-75-5
Access Level:Open access
Keyword:Extractive question answering
Cross-lingual transfer
Knowledge distillation
Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Llenguatge natural
Description
Summary:We address the challenge of Generalized Cross-Lingual Transfer (G-XLT) in extractive Question Answering, where question and context languages differ, a problem particularly difficult for low-resource languages. Working with only a thousand parallel QA samples, we combine cross-lingual sampling with self-knowledge distillation to regularize cross-lingual fine-tuning. We introduce the novel mean Average Precision at k (mAP@k) coefficient, which mitigates the negative impact of incorrect predictions during training and serves as a diagnostic tool providing early training guidance and reliable indicators of model learning. Evaluations on MLQA, XQuAD, and TyDiQA-GoldP datasets demonstrate that our approach consistently outperforms standard cross-entropy fine-tuning of the mBERT multilingual model. Our method represents a promising alternative to machine translation-based approaches, particularly valuable for low-resource languages where translation quality is poor, offering an efficient solution for cross-lingual transfer in data-scarce settings.