SOURCE-LINKED INTELLIGENCE
RePair: Turning Retrieval Failures into Counterfactual Hard Pairs
Vision-language retrieval with CLIP-style dual encoders achieves strong cross-modal performance, yet practical accuracy often hinges on localized semantic distinctions where top-ranked near misses differ from the true match by a single critical detail. Hard-sample mining can select confusable candidates but cannot construct corrected counterparts; synthetic augmentation can generate novel samples but, without conditioning on actual model failures, targets irrelevant dimensions of hardness. We observe that a top-ranked false positive is a counterfactual scaffold---sharing most of the query's se
Read original source ↗ Open in workspace
- recordType
- paper
- region
- Global
Evidence & attribution
- arXiv · AI, language, vision and robotics · 2026-08-30T06:47:38.000Z
First collected: 2026-09-21T07:31:56.984Z. This is not the publication date.