Metraipolis › archive › dc_bfd8be46beb6
Align, Integrate, and Fire: Efficient Token-Level Alignment for Zero-Shot SpeechLLMs — arXiv:2609.18516
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.18516
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T13:49:28.970Z
- SHA-256 of the archived text
f9e9c3b1032804e7a4600398b211d1e3fe71dff57c1ce1c66576f7d804633785- Stable URL of this page
https://metraipolis.com/documento/dc_bfd8be46beb6
How to cite
Align, Integrate, and Fire: Efficient Token-Level Alignment for Zero-Shot SpeechLLMs — arXiv:2609.18516 — arXiv (en), https://arxiv.org/abs/2609.18516 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bfd8be46beb6 (Metraipolis archive dc_bfd8be46beb6, sha256 f9e9c3b1032804e7a4600398b211d1e3fe71dff57c1ce1c66576f7d804633785)
BibTeX
@misc{dcbfd8be46beb6,
title = {Align, Integrate, and Fire: Efficient Token-Level Alignment for Zero-Shot SpeechLLMs — arXiv:2609.18516},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.18516},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bfd8be46beb6; sha256 f9e9c3b1032804e7a4600398b211d1e3fe71dff57c1ce1c66576f7d804633785},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.18516. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
While Large Language Models excel in natural language processing, efficiently extending their capabilities to spoken input remains a significant challenge. Existing methods for building SpeechLLMs often rely on computationally expensive full-model fine-tuning, or employ parameter-efficient projectors that suffer from inefficient token sequence lengths and costly full-model supervision. In this paper, we introduce Aligned Continuous Integrate-and-Fire, a highly efficient framework for zero-shot speech processing. Our method dynamically compresses continuous acoustic frames into the exact discrete token length of the target text utilizing explicit Dynamic Time Warping alignments. This allows our initial training stage to establish a robust acoustic-to-semantic bridge using lightweight distance metrics, entirely bypassing the computationally expensive LLM forward pass. For subsequent fine-tuning, we propose a memory-efficient knowledge distillation objective that targets a single LLM layer, performing competitively with full-model cross-entropy training at a fraction of the computational cost. Through extensive evaluations on Automatic Speech Recognition and Speech Translation, we demonstrate that our method achieves superior performance compared to prior parameter-efficient baselines. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-17 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.18516 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-15T13:48:24.324Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).