LVLM unlearning benchmarks fail due to initial memorization failures on fictitious data; ReMem benchmark with multi-hop and multi-image scaling plus Exposure metric enables reliable learning and unlearning diagnosis.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
DiffRGD is a plug-and-play inference-time guidance method that casts each diffusion sampling step as constrained optimization on a spherical manifold and solves it with Riemannian gradient descent to preserve the Gaussian latent structure.
citing papers explorer
-
Before Forgetting, Learn to Remember: Revisiting Foundational Learning Failures in LVLM Unlearning Benchmarks
LVLM unlearning benchmarks fail due to initial memorization failures on fictitious data; ReMem benchmark with multi-hop and multi-image scaling plus Exposure metric enables reliable learning and unlearning diagnosis.
-
DiffRGD: An Inference-Time Diffusion Guidance Through Riemannian Gradient Descent
DiffRGD is a plug-and-play inference-time guidance method that casts each diffusion sampling step as constrained optimization on a spherical manifold and solves it with Riemannian gradient descent to preserve the Gaussian latent structure.