Tag: preference-optimization
All the articles with the tag "preference-optimization".
- 5.0
Rethinking Efficiency in Neural Combinatorial Optimization: Batched Preference Optimization with Mamba
ECO 将 NCO 效率作为一等目标,结合批量化 PO 与 Mamba backbone 解耦轨迹生成与梯度更新
- 5.5
Intrinsic Mutual Information as a Modulator for Preference Optimization
用互信息作为调制因子自适应调整 DPO 超参,减少手动调优时间开销