DPO with contrastive sequence-annotation alignment improves GO term prediction by 2 to 4 percent relative F1-Max over supervised fine-tuning alone.
Aligning protein generative models with experimental fitness via direct preference optimization
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
q-bio.BM 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
AnnoDPO: Protein Functional Annotation Learning with Direct Preference Optimization
DPO with contrastive sequence-annotation alignment improves GO term prediction by 2 to 4 percent relative F1-Max over supervised fine-tuning alone.