DANCE通过双层自适应邻域法，提升预训练模型的置信集效率与鲁棒性。
DANCE: Doubly Adaptive Neighborhood Conformal Estimation
- 基于嵌入表示构建双层自适应邻域，融合两个新颖非符合度评分
- 在多个数据集上实现更小置信集且保持统计有效性
- 适合需要高精度不确定性量化任务的机器学习应用
复杂深度学习模型在多种数据表示类型上的预测能力已达到前所未有的水平。为量化这些模型的不确定性，适配性、统计有效的分位数预测方法日益流行。传统分类任务中的分位数方法通常依赖对数几率分数，但对于预训练模型而言，若未针对目标任务校准，可能导致置信集过大、过于保守。本文提出DANCE：一种基于双重局部自适应最近邻的分位数算法，直接利用数据嵌入表示构造两个新颖的非符合度评分。DANCE首先在嵌入层拟合任务自适应核回归模型，再在学习到的核空间中生成最终预测集以实现不确定性量化。我们在多个数据集上对比了最先进的局部化、任务自适应及零样本分位数基线方法，结果表明DANCE在置信集大小效率与鲁棒性之间表现出更优平衡。
原文摘要 · Abstract (English)
The recent developments of complex deep learning models have led to unprecedented ability to accurately predict across multiple data representation types. Conformal prediction for uncertainty quantification of these models has risen in popularity, providing adaptive, statistically-valid prediction sets. For classification tasks, conformal methods have typically focused on utilizing logit scores. For pre-trained models, however, this can result in inefficient, overly conservative set sizes when not calibrated towards the target task. We propose DANCE, a doubly locally adaptive nearest-neighbor based conformal algorithm combining two novel nonconformity scores directly using the data's embedded representation. DANCE first fits a task-adaptive kernel regression model from the embedding layer before using the learned kernel space to produce the final prediction sets for uncertainty quantification. We test against state-of-the-art local, task-adapted and zero-shot conformal baselines, demonstrating DANCE's superior blend of set size efficiency and robustness across various datasets.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。