arXiv:2609.21061v1 Announce Type: new Abstract: 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 u

So what

What this event means by reading role—not a longer recap.

BuilderIf you fine-tune vision models on scarce, domain-shifted imagery, a single LoRA stage at rank 4 (0.26% of weights) may beat stacked contrastive refinement pipel
ResearcherThe paper's null results suggest hard-negative mining and SupCon add nothing once the encoder has already fit the data, a useful negative control for parameter-

Read original Back to Stories

Score dimensions

Higher total means read first. Each bar is one factor we use to rank the system pool. How we score

RelevanceHow tightly this is about AI.: 62
ImpactHow much this could change the field or the market.: 35
NoveltyHow new this is versus a recap.: 55
CredibilityHow much we trust the source.: 92
ActionabilityWhether a reader can do something with it.: 30