New cycle-consistent optimization, task vector theory, singular vector decompositions, adaptive routing, and efficient evolutionary search provide foundations for merging neural network weights across tasks.
Title resolution pending
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
dataset 1
citation-polarity summary
years
2026 3roles
dataset 1polarities
use dataset 1representative citing papers
Repurposing discarded supervised classification heads from vision models as semantic prototypes boosts post-hoc vision-language alignment methods on retrieval and zero/few-shot classification.
citing papers explorer
-
Model Merging: Foundations and Algorithms
New cycle-consistent optimization, task vector theory, singular vector decompositions, adaptive routing, and efficient evolutionary search provide foundations for merging neural network weights across tasks.
-
Supervised Classification Heads as Semantic Prototypes: Unlocking Vision-Language Alignment via Weight Recycling
Repurposing discarded supervised classification heads from vision models as semantic prototypes boosts post-hoc vision-language alignment methods on retrieval and zero/few-shot classification.
- Unlocking the Potential of Continual Model Merging: An ODE Perspective