This is the official implementation of Understanding Narratives through Dimensions of Analogy (the camera-ready version of the paper is HERE) which has been accepted by the HERE.
Please see our paper for more information and discussion.
git clone https://github.com/usc-isi-i2/analogical-transfer-learning
cd analogical-transfer-learning
pip install -e .
- Scrape Dataset.ipynb - Data collection script
- Cluster Fables.pynb - Task 1: Moral Clustering | Language model baseline
- Frames Based Clustering.ipynb - Task 1: Moral Clustering | Frame model baseline
- Emotional Arc.ipynb - Task 2: Analogical Pair Generation | Story shape baseline
- Story Pairs.ipynb - Task 3: Analogy Type Prediction
- Transfer Learning.pynb - Task 4: Analogical Transfer
The relavent data that is generated or used in the notebooks are available here
All details on this project are documented here and the related files are stored here
Updated analogy language model code can be found here