Adversarial Robustness of Graph and Vision-Language Contrastive Learning on Multimodal Benchmarks

This report synthesises findings from 9 peer-reviewed papers addressing the following research question: How does the adversarial robustness of graph contrastive learning methods compare to vision-language contrastive models when evaluated on multimodal reasoning benchmarks under similar perturbation. In the last few years, the deep learning (DL) computing paradigm has been deemed the Gold Standard in the machine learning (ML) community. Moreover, it has gradually become the most widely used computational approach in the field of ML, thus achieving outstanding results on. 9 claims were extracted from source literature; 9 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 8.7/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: How does the adversarial robustness of graph contrastive learning methods compare to vision-language contrastive models when evaluated on multimodal reasoning benchmarks under similar perturbation levels?

Autonomous literature synthesis. Automated review score: 8.7/10. Full text and citation available at Assignee Research.