Questions for Active Reading

Code and Data for How to Engage Your Reader? Generating Guiding Questions to Promote Active Reading

Peng Cui, Xiaoyu Zhang, Vilém Zouhar, and Mrinmaya Sachan.

---

Introduction

We study how human writers use questions in academic writing and how these questions influence human reading comprehension. To this end, we

1. curate GuidingQ, a dataset of in-text questions in textbooks from Openstax and research articles from arXiv and Pubmed.
2. fine-tune various LMs to generate these questions to promote active reading.
3. conduct a human study to investigate the effect of such questions.

---

Dataset

Our dataset is available at https://drive.google.com/file/d/1BlMUkQAt95pL8V4Mj_Y9I-X8FPCiWNl9/view?usp=drive_link.

Each example consists of

{
    "id": "article_id",
    "title": "article_title",
    "article": "list of article sentences",
    "questions": "list of extracted questions and their information"
}

---

Experiment

• Code for model training and evaluation: src.
• Usage: src/main.
• Model ckpt: finetuned models.

---

Human Study

We release our human study UI for similar research in the future. See a demo version.

Screen pipeline:

1. Demographic questions
2. Intructions
3. Reading task
4. Performance questions
5. Project questions
6. Reading interface (again) with question annotation
7. Exit

Running this Experiment Interface

First you need to start the logger locally:

git clone https://github.com/zouharvi/annotation-logger.git
cd annotation-logger
# run it in the background
nohup python3 main.py &

Then, build

git clone https://github.com/zouharvi/reading-comprehension-help
cd reading-comprehension-help/annotation_ui
npm install
# run locally and use local logger, not the live one
npm run dev
# go to this url which loads the file `annotation_ui/web/queues/demo_authentic.jsonl`
xdg-open localhost:9001?uid=demo_authentic

You can go from here by adding new user queues to annotation_ui/web/queues and loading them with the uid= parameter.

Cite

@inproceedings{cui-etal-2024-engage,
    title = "How to Engage your Readers? Generating Guiding Questions to Promote Active Reading",
    author = "Cui, Peng  and
      Zouhar, Vil{\'e}m  and
      Zhang, Xiaoyu  and
      Sachan, Mrinmaya",
    editor = "Ku, Lun-Wei  and
      Martins, Andre  and
      Srikumar, Vivek",
    booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.acl-long.632",
    pages = "11749--11765"
}

Contact

For any questions, feel free to open an issue or drop me an email at peng.cui@inf.ethz.ch.