Controlled vision experiments show that linear weight merging preserves shared knowledge but rapidly erases task-specific (unshared) knowledge, and merging sequentially trained models is safer than merging parallel-trained ones.
Where is the truth? the risk of getting confounded in a continual world
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
Forgetting of task-specific knowledge in model merging-based continual learning
Controlled vision experiments show that linear weight merging preserves shared knowledge but rapidly erases task-specific (unshared) knowledge, and merging sequentially trained models is safer than merging parallel-trained ones.