TILT通过目标端惩罚提升域适应性能，解决分布偏移问题。
TILT: Target-induced loss tilting under covariate shift
- 将源域预测器分解为两部分，对目标域无标签数据施加辅助项惩罚
- 在有限样本下实现误差上界控制，支持稀疏ReLU网络训练
- 适用于分布偏移场景，对正则化参数不敏感，性能稳定
我们提出针对协变量偏移下无监督域适应的靶向损失倾斜（TILT）方法。其核心是新型目标函数：将源域预测器分解为 $f+b$，在有标签源数据上拟合 $f+b$，同时在无标签目标数据上对辅助成分 $b$ 施加惩罚，最终部署 $f$ 作为目标域预测器。在总体层面，该目标侧惩罚隐式诱导了相对重要性加权，但针对当前误差自定位的估计量 $b^*_f$，且对任意源-目标对（包括支撑集不重叠情况）均保持一致有界。我们建立了通用的有限样本贝叶斯风险上界，并据此给出稀疏ReLU网络训练的端到端保证。在受控回归问题和偏移后的CIFAR-100蒸馏实验中，TILT优于仅源域训练、精确重要性加权和相对密度比基线，且对正则化参数具有稳定依赖性。
原文摘要 · Abstract (English)
We introduce and analyze Target-Induced Loss Tilting (TILT) for unsupervised domain adaptation under covariate shift. It is based on a novel objective function that decomposes the source predictor as $f+b$, fits $f+b$ on labeled source data while simultaneously penalizing the auxiliary component $b$ on unlabeled target inputs. The resulting fit $f$ is deployed as the final target predictor. At the population level, we show that this target-side penalty implicitly induces relative importance weighting at the population level, but in terms of an estimand $b^*_f$ that is self-localized to the current error, and remains uniformly bounded for any source-target pair (even those with disjoint supports). We prove a general finite-sample oracle inequality on the excess risk, and use it to give an end-to-end guarantee for training with sparse ReLU networks. Experiments on controlled regression problems and shifted CIFAR-100 distillation show that TILT improves target-domain performance over source-only training, exact importance weighting, and relative density-ratio baselines, with a stable dependence on the regularization parameter.
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。