Story Embeddings

This repository contains the code for both the papers "Tell Me Again! a Large-Scale Dataset of Multiple Summaries for the Same Story" and "Story Embeddings — Narrative-Focused Representations of Fictional Stories".

The code is licensed under MIT.

When using any of our code, we kindly ask that you reference our paper:

@inproceedings{hatzel-biemann-2024-story-embeddings,
    title = "Story Embeddings -- Narrative-Focused Representations of Fictional Stories",
    author = "Hatzel, Hans Ole and Biemann, Chris",
    booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics",
    year = "2024",
    address = "Miami, Florida, USA",
    publisher = "Association for Computational Linguistics",
}

Quick Overview

The most important things you could be looking for are listed below:

• Contrastive learning for story embeddings: embeddings.py
• Retelling Dataset: data/retellings/
• Dataset scraping: main.py unless you know what you are doing, just pip install tell-me-again instead!

Tell Me Again Data

You shouldn't need to scrape the Tell Me Again data yourself (unless you want to build a new version of the dataset). Instead just use the package tell-me-again.

Models

Our model trained with augmented data is available on huggingface.