Metraipolis › archive › dc_bbda71869c81

LoRA Enhanced Contrastive Learning with SAS Vision Transformers — arXiv:2609.21061

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.21061
License (as archived): CC0-1.0
Archived: 2026-10-10T01:01:42.917Z
SHA-256 of the archived text: 6320f008a3262cbb3485287934687c9de778d699257937c46b71b08ae8630ef3
Stable URL of this page: https://metraipolis.com/documento/dc_bbda71869c81

How to cite

LoRA Enhanced Contrastive Learning with SAS Vision Transformers — arXiv:2609.21061 — arXiv (en), https://arxiv.org/abs/2609.21061 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bbda71869c81 (Metraipolis archive dc_bbda71869c81, sha256 6320f008a3262cbb3485287934687c9de778d699257937c46b71b08ae8630ef3)

BibTeX

@misc{dcbbda71869c81,
  title = {LoRA Enhanced Contrastive Learning with SAS Vision Transformers — arXiv:2609.21061},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.21061},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bbda71869c81; sha256 6320f008a3262cbb3485287934687c9de778d699257937c46b71b08ae8630ef3},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.21061. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Automatic target recognition (ATR) with synthetic aperture sonar (SAS) supports advanced naval capabilities, but deep learning is constrained by scarce target imagery, background clutter, and human-in-the-loop assessment. We adapt DINOv3 Vision Transformer (ViT) models to underwater SAS ATR using a three-stage parameter-efficient framework. Stage 1 uses Low-Rank Adaptation (LoRA) while freezing the ViT backbone, bridging the gap between natural-image pretraining and underwater acoustic propagation. Stage 2 uses hard-negative mining to strengthen the decision boundary against acoustic mimics, including rocks and sediment formations resembling man-made targets. Stage 3 uses Supervised Contrastive Learning (SupCon) to separate target and clutter representations. We evaluate at-sea SAS data using a mission-level geographic split, compare all arms at 85 percent test recall, and repeat each comparison over three random seeds. LoRA accounts for the primary effect, increasing area under the precision-recall curve (AUPRC) from 0.300 to 0.679 +/- 0.027 using the same frozen backbone. Rank 4 achieves this result while training only 0.26 percent of weights. Neither refinement stage exceeds its matched control: hard-negative mining changes AUPRC by -0.0045 +/- 0.0119 versus an equal-size random curriculum, and SupCon changes AUPRC by +0.0002 +/- 0.0096 versus the preceding stage. These null results indicate that mining occurred on data the encoder had already fit and that supervised stages had already imposed most target-clutter geometry. One efficient adaptation stage is sufficient; stacked refinement is not.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-21 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.21061 — 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 2026-10-10T00:23:42.426Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-21?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).