TIDE Paper Speeds Up Diffusion Training Data Attribution by Orders of Magnitude

Sony and Adobe researchers introduce TIDE, which distills gradient-based attribution into low-dimensional embeddings, enabling millisecond training data attribution for diffusion models—4-5 orders of magnitude faster than before.

2026-10-09 ~ 2026-10-09 · 3 related posts

• TIDE attributes diffusion outputs to training images in milliseconds — serrjoa · 2026-10-09
• Distilled influence embeddings make diffusion training data attribution a nearest-neighbor lookup — serrjoa · 2026-10-09

1 near-duplicate retellings: serrjoa