@inproceedings{kolluri-etal-2026-role,
    title = "On the Role of Encoder Depth: Pruning Whisper and {L}o{RA} Fine-Tuning in {SLAM}-{ASR}",
    author = "Kolluri, Ganesh Pavan Kartikeya Bharadwaj  and
      Kampouridis, Michael  and
      Shekhar, Ravi",
    editor = "Hosseini-Kivanani, Nina  and
      Brutti, Alessio  and
      Matassoni, Marco  and
      Dowerah, Sandipana  and
      Liga, Davide  and
      Schommer, Christoph",
    booktitle = "Proceedings of Speech Language Models in Low-Resource Settings: Performance, Evaluation, and Bias Analysis ({SPEAKABLE}) @ {LREC} 2026",
    month = may,
    year = "2026",
    address = "Palma, Mallorca (Spain)",
    publisher = "ELRA Language Resources Association (ELRA)",
    url = "https://preview.aclanthology.org/revision-previews/2026.speakable-1.20/",
    doi = "10.63317/2pbwm223zmqx",
    pages = "183--193",
    abstract = "Automatic speech recognition (ASR) has advanced rapidly in recent years, driven by large-scale pretrained models and end-to-end architectures such as SLAM-ASR. A key component of SLAM-ASR systems is the Whisper speech encoder, which provides robust acoustic representations. While model pruning has been explored for the full Whisper encoder{--}decoder architecture, its impact within the SLAM-ASR setting remains under-investigated. In this work, we analyze the effects of layer pruning in the Whisper encoder when used as the acoustic backbone of SLAM-ASR. We further examine the extent to which LoRA-based fine-tuning can recover performance degradation caused by pruning. Experiments conducted across three Whisper variants (Small, Medium, Large-v2), three languages representing distinct resource levels (Danish, Dutch, English), and over 200 training runs demonstrate that pruning two encoder layers causes only 2{--}4{\%} WER degradation, and that combining this pruning with LoRA adaptation consistently outperforms the unpruned baseline while reducing total parameters by 7{--}14{\%}. Moreover, our error analysis reveals that LoRA primarily compensates through the language model{'}s linguistic priors, reducing total word errors by 18.2{\%}, with substitution errors showing the largest reduction. However, for low-resource Danish, LoRA introduces increased insertion errors, indicating that compensation effectiveness depends on the LLM{'}s pre-existing language proficiency and available training data."
}