@inproceedings{rao-etal-2025-whispa,
    title = "{W}hi{SPA}: Semantically and Psychologically Aligned Whisper with Self-Supervised Contrastive and Student-Teacher Learning",
    author = "Rao, Rajath  and
      V Ganesan, Adithya  and
      Kjell, Oscar  and
      Luby, Jonah  and
      Raghavan, Akshay  and
      Feltman, Scott M.  and
      Ringwald, Whitney  and
      Boyd, Ryan L.  and
      Luft, Benjamin J.  and
      Ruggero, Camilo J.  and
      Ryant, Neville  and
      Kotov, Roman  and
      Schwartz, H.",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingestion-acl-25/2025.acl-long.1098/",
    pages = "22529--22544",
    ISBN = "979-8-89176-251-0",
    abstract = "Current speech encoding pipelines often rely on an additional text-based LM to get robust representations of human communication, even though SotA speech-to-text models often have a LM within. This work proposes an approach to improve the LM within an audio model such that the subsequent text-LM is unnecessary. We introduce **WhiSPA** (**Whi**sper with **S**emantic and **P**sychological **A**lignment), which leverages a novel audio training objective: contrastive loss with a language model embedding as a teacher. Using over 500k speech segments from mental health audio interviews, we evaluate the utility of aligning Whisper{'}s latent space with semantic representations from a text autoencoder (SBERT) and lexically derived embeddings of basic psychological dimensions: emotion and personality. Over self-supervised affective tasks and downstream psychological tasks, WhiSPA surpasses current speech encoders, achieving an average error reduction of 73.4{\%} and 83.8{\%}, respectively. WhiSPA demonstrates that it is not always necessary to run a subsequent text LM on speech-to-text output in order to get a rich psychological representation of human communication."
}