Tag
该论文研究 LoRA 适配器初始输入子空间方向的重要性为何依赖于数据 regime,提出基于激活 PCA 的 PivotLoRA 初始化方法,并在少样本 BERT/Llama 分类及 GSM8K、StrategyQA 任务上验证其优于原版 LoRA。
This paper explores activation steering for automatic speech recognition (ASR), finding that supervised optimization of steering vectors improves both overall ASR performance and demographic fairness across gender, age, and dialect, while also enabling few-shot or zero-shot adaptation using acoustically similar speakers.
The paper proposes Approximate Pareto Optimality (APO), a method for collaboratively learning lightweight LLM aligner initializations that support few-shot personalization under data deficiency, grouping users by compatible updates and coordinating competing objectives toward Pareto-optimal points with proven suboptimality bounds.
MC-RFM proposes a novel Riemannian flow-matching framework for few-shot adaptation that models feature displacement on a mixed-curvature manifold combining hyperbolic and Euclidean spaces, outperforming existing methods across multiple visual recognition benchmarks.