Pre-trained models have shown very good performances on a number of question answering benchmarks especially when fine-tuned on multiple question answering datasets at once. In this work, we generate a fine-tuning dataset model thanks to a rule-based algorithm that generates questions and answers from unannotated sentences. We show that the state-of-the-art model UnifiedQA can greatly benefit from such a system on a multiple-choice benchmark about physics, biology and chemistry it has never been trained on. We further show that improved performances may be obtained by selecting the most challenging distractors (wrong answers), with a dedicated ranker based on a pretrained RoBERTa model.

La présentation sera donnée en français.

Enregistrement vidéo: https://drive.google.com/file/d/11dBS-RL8zNON9RGsniQbUkbuFV_9WoCe/view?usp=sharing