Lluna: graph-based unimodal collaboration fixes missing-modality inference failures
williammary · hf · 2026-10-16
Researchers from William & Mary, CMU, Microsoft and UChicago show that multimodal collaboration among unimodal models can weaken or even reverse when modalities go missing. Lluna combines a direct private prediction path from adapted unimodal features with a graph-based residual collaboration branch, plus Progressive Collaboration Consistency training. Across 8 datasets spanning affective computing, scene understanding and medical prediction, Lluna achieves the best average performance evaluated.
More from Multimodal
- DLSS 5 Assetto Corsa demo stuns sim racers: 'better than reality' — CSProfKGD · 2026-10-16
- Video recast with fal agent: swap a character via one character sheet — adamho · 2026-10-16
- Opus 5.5 + Midjourney: 'models have taste, you just point them right' — yuwen_lu_ · 2026-10-16
- Gemini 4 Argon hits AssemblyAI's LLM Gateway, tops voice bench at 99.5 across 50+ models — AssemblyAI · 2026-10-16
- Envato Launches AI Creative Toolkit Bundling Nano Banana, Veo and More — alifcoder · 2026-10-16
- Musk amplifies Grok Imagine dance video demo tackling generative AI's hardest motion — elonmusk · 2026-10-16