@inproceedings{fan-etal-2025-medium,
    title = "The Medium Is Not the Message: Deconfounding Document Embeddings via Linear Concept Erasure",
    author = "Fan, Yu  and
      Tian, Yang  and
      Ravfogel, Shauli  and
      Sachan, Mrinmaya  and
      Ash, Elliott  and
      Hoyle, Alexander Miserlis",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2025.emnlp-main.1634/",
    pages = "32112--32131",
    ISBN = "979-8-89176-332-6",
    abstract = "Embedding-based similarity metrics between text sequences can be influenced not just by the content dimensions we most care about, but can also be biased by spurious attributes like the text{'}s source or language. These document confounders cause problems for many applications, but especially those that need to pool texts from different corpora. This paper shows that a debiasing algorithm that removes information about observed confounders from the encoder representations substantially reduces these biases at a minimal computational cost. Document similarity and clustering metrics improve across every embedding variant and task we evaluate{---}often dramatically. Interestingly, performance on out-of-distribution benchmarks is not impacted, indicating that the embeddings are not otherwise degraded."
}