Abstract

This study addressed the challenge of devising a language model pipeline proficient in generating tailored quizzes comprising multiple question-answer pairs on specific themes. Recent advancements in natural language processing showcase the capabilities of language models in generating coherent and contextually relevant text [1][2][3]. Leveraging this progress, the focus was on constructing a dynamic pipeline integrating multiple language models to facilitate the generation of diverse and accurate quiz content. The primary objective was to design a system that amalgamates the strengths of various language models, optimizing their collective abilities to generate comprehensive and educative quiz sets. This project was carried out as part of the Python programming course at ENSAE Paris. It addresses data processing, data visualization and modeling works.

Collaborators

This work has been done in a group work with Adrien Servière , under the supervision of Kim Antunez from INSEE as part of the Python Programming course at ENSAE Paris.

References

[1] Radford, A., Wu, J., Child, R., Luan, D., Amodei, D., & Sutskever, I. (2019). Language Models are Unsupervised Multitask Learners.

[2] Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2019). Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer.

[3] Keskar, N. S., McCann, B., Varshney, L. R., Xiong, C., & Socher, R. (2019). CTRL: A Conditional Transformer Language Model for Controllable Generation.