Benchmark RadarAI BENCHMARK PROFILE
MMDG-Bench
MMDG-Bench is a benchmark for multi-modal domain generalization, providing two frameworks (D2M and M2D) and unified protocols across action recognition and face anti-spoofing tasks. Includes code repository.
- Released
- 2026-05-30
- Readiness
- Runnable
- Primary field
- General AI
Why it matters
Provides a principled foundation for MMDG research, with insights on framework choice and backbone effects, and offers design guidelines for multi-modal robustness.
Motivation
Multi-modal Domain Generalization (MMDG) seeks to leverage complementary modalities to enhance model robustness on unseen domains.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.