Hard, Yet Reducible: Controlled Forward Transfer for Synthetic Degradation Curation
We propose the controlled Reducible Degradation Gap (cRDG) for regions defined by degradation type and severity.
ProofPaper ↗
Key points
- Selecting synthetic degradations for dense prediction requires an estimate of their training utility, the generalization gain they bring under a finite training budget.
- Clean and degraded twins share content and labels, suggesting a score based on how much short training reduces the excess error caused by degradation.
- Curation of Reducible Bands (\method) uses cRDG to select synthetic data without changing the predictor.
- On semantic segmentation and salient object detection, \method{} improves representative predictors under matched synthetic-data budgets and training schedules, extends to existing data-generation pipelines, and preserves clean performance.
Sources (1)
- [1]Hard, Yet Reducible: Controlled Forward Transfer for Synthetic Degradation CurationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 11:07 AM
We propose the controlled Reducible Degradation Gap (cRDG) for regions defined by degradation type and severity.
Selecting synthetic degradations for dense prediction requires an estimate of their training utility, the generalization gain they bring under a finite training budget.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 6, 2026World Models Dream of Success: Diagnosing and Repairing Failure Insensitivity in Robot World Models
- Oct 6, 2026Learning Transition Kernels of Jump-Diffusion Processes with Conditional Diffusion Models
- Oct 6, 2026Algorithmic Scratchpads and Curriculum Staging for Arithmetic Reasoning in Tiny Transformers
- Oct 6, 2026Build a voice travel concierge with Amazon Bedrock AgentCore, Managed Knowledge Base and Nova Sonic
- Oct 4, 2026SheetSage2: Coherent Lead-Sheet Transcription with Synthetic Supervision
- Oct 2, 2026Inside-Out AI: Rebuilding Airbnb Behind the Scenes and Across the Guest Experience