Metraipolis › archive › dc_d7efedcb4e27

Instruction-Conditioned Electromagnetic Spectrum Understanding via Budget-Adaptive Signal Tokenization — arXiv:2610.12142

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.12142
License (as archived): CC0-1.0
Archived: 2026-10-09T05:50:33.990Z
SHA-256 of the archived text: bf99c0621dc36d6ca72b529f4fe37c3a4d8f26d70a07a791ef5caab050649c42
Stable URL of this page: https://metraipolis.com/documento/dc_d7efedcb4e27

How to cite

Instruction-Conditioned Electromagnetic Spectrum Understanding via Budget-Adaptive Signal Tokenization — arXiv:2610.12142 — arXiv (en), https://arxiv.org/abs/2610.12142 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d7efedcb4e27 (Metraipolis archive dc_d7efedcb4e27, sha256 bf99c0621dc36d6ca72b529f4fe37c3a4d8f26d70a07a791ef5caab050649c42)

BibTeX

@misc{dcd7efedcb4e27,
  title = {Instruction-Conditioned Electromagnetic Spectrum Understanding via Budget-Adaptive Signal Tokenization — arXiv:2610.12142},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.12142},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d7efedcb4e27; sha256 bf99c0621dc36d6ca72b529f4fe37c3a4d8f26d70a07a791ef5caab050649c42},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.12142. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Electromagnetic spectrum monitoring increasingly requires flexible analysis beyond task-specific recognition and detection. Multimodal large language models offer a unified interface, but extending vision-language models (VLMs) to raw I/Q signals requires tokenization that balances fidelity against a strict budget. For signals, dense encoding causes token costs to grow with observation length, whereas fixed-resolution compression may discard short-duration or localized signal evidence. Thus, we propose \textbf{BATok}, a budget-adaptive signal tokenizer that adjusts token capacity to the input length while allocating that capacity according to the signal content. BATok constructs candidate representations from signal-derived features using lightweight multi-resolution branches, then combines a local energy prior with learnable queries to resample these representations into compact signal tokens. The number of tokens adapts to the input length while remaining strictly bounded. The resulting tokens are projected into the language embedding space of VLMs. We further introduce \textbf{EMSpec-Instruct}, a multimodal instruction dataset aligning raw I/Q signals, waterfall images, and language supervision for modulation recognition, structured detection, and language-conditioned signal grounding. Experiments show that BATok learns effective signal representations and achieves competitive performance across all tasks.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Signal Processing (eess.SP).
Pagina del lavoro: https://arxiv.org/abs/2610.12142 — 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 2026-10-09T05:47:01.087Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).