Learning Task Mixtures from Task Affinities: A Probabilistic Graphical Model for Supervised Fine-Tuning
This work introduces TaskPGM, a framework for learning continuous task mixtures via an energy-based model over tasks, and shows that the resulting set function is weakly submodular under budget constraints, enabling approximation guarantees for discrete selection variants.
Prateek Chanda, Saral Sureka, Parthiv Chatterjee et al.
· 1 citation