Isolating authorship from content with semantic embeddings and contrastive learning
This work addresses authorship attribution for forensic or literary analysis, but it is incremental as it builds on existing contrastive learning methods.
The paper tackled the problem of disentangling authorship style from content in text by using contrastive learning with synthetic hard negatives from a semantic similarity model, resulting in up to a 10% accuracy increase in challenging authorship attribution tasks.
Authorship has entangled style and content inside. Authors frequently write about the same topics in the same style, so when different authors write about the exact same topic the easiest way out to distinguish them is by understanding the nuances of their style. Modern neural models for authorship can pick up these features using contrastive learning, however, some amount of content leakage is always present. Our aim is to reduce the inevitable impact and correlation between content and authorship. We present a technique to use contrastive learning (InfoNCE) with additional hard negatives synthetically created using a semantic similarity model. This disentanglement technique aims to distance the content embedding space from the style embedding space, leading to embeddings more informed by style. We demonstrate the performance with ablations on two different datasets and compare them on out-of-domain challenges. Improvements are clearly shown on challenging evaluations on prolific authors with up to a 10% increase in accuracy when the settings are particularly hard. Trials on challenges also demonstrate the preservation of zero-shot capabilities of this method as fine tuning.