Minimizing contrastive loss produces an orthogonal modality gap vector whose size is monotonically tied to robustness, so post-processing that reduces the gap improves robustness with no loss in clean accuracy.
Imagenet: A large-scale hierarchical image database
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3verdicts
UNVERDICTED 3representative citing papers
Synthetic facial images alone can train models for pediatric rare disease recognition to performance levels comparable to real-data baselines when generated at sufficient scale.
DRIFT uses a multi-frame network trained with adversarial perceptual loss for alignment, denoising, demosaicing and super-resolution, followed by an efficient deep tone-mapping module that supports tunability and reference consistency.
citing papers explorer
-
Is the Modality Gap a Bug or a Feature? A Robustness Perspective
Minimizing contrastive loss produces an orthogonal modality gap vector whose size is monotonically tied to robustness, so post-processing that reduces the gap improves robustness with no loss in clean accuracy.
-
Synthetic Data Alone is Enough? Rethinking Data Scarcity in Pediatric Rare Disease Recognition
Synthetic facial images alone can train models for pediatric rare disease recognition to performance levels comparable to real-data baselines when generated at sufficient scale.
-
DRIFT: Deep Restoration, ISP Fusion, and Tone-mapping
DRIFT uses a multi-frame network trained with adversarial perceptual loss for alignment, denoising, demosaicing and super-resolution, followed by an efficient deep tone-mapping module that supports tunability and reference consistency.