An energy-reweighted flow-matching loss is shown by construction to target a reweighted distribution, with modest benchmark demonstrations of concept removal.
The right panel illustrates the learned and unlearned trajectories for four 2D benchmarks: Circles, Moons, 6 Gaussians, and Checkerboard
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
ContinualFlow: Learning and Unlearning with Neural Flow Matching
An energy-reweighted flow-matching loss is shown by construction to target a reweighted distribution, with modest benchmark demonstrations of concept removal.