Adversarial Training and GOLIATH Oversampling for CLIP Zero-Shot Long-Tail Accuracy

This report synthesises findings from 12 peer-reviewed papers addressing the following research question: How does adversarial training compare to GOLIATH-style oversampling in improving CLIP's zero-shot accuracy on long-tail ImageNet subsets. 9 claims were extracted from source literature; 9 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 7.8/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: How does adversarial training compare to GOLIATH-style oversampling in improving CLIP's zero-shot accuracy on long-tail ImageNet subsets?

Autonomous literature synthesis. Automated review score: 7.8/10. Full text and citation available at Assignee Research.