Abstract
The learning order of semantic classes significantly impacts unsupervised domain adaptation for semantic segmentation, especially under adverse weather conditions. Most existing curricula rely on handcrafted heuristics (e.g., fixed uncertainty metrics) and follow a static schedule, which fails to adapt to a model's evolving, high-dimensional training dynamics, leading to category bias. Inspired by Reinforcement Learning, we cast curriculum learning as a sequential decision problem and propose an autonomous class scheduler. This scheduler consists of two components: (i) a high-dimensional state encoder that maps the model's training status into a latent space and distills key features indicative of progress, and (ii) a category-fair policy-gradient objective that ensures balanced improvement across classes. Coupled with mixed source-target supervision, the learned class rankings direct the network's focus to the most informative classes at each stage, enabling more adaptive and dynamic learning. It is worth noting that our method achieves state-of-the-art performance on three widely used benchmarks (e.g., ACDC, Dark Zurich, and Nighttime Driving) and shows generalization ability in synthetic-to-real semantic segmentation.
Abstract (translated)
语义类别的学习顺序显著影响无监督领域适应在语义分割中的表现，尤其在恶劣天气条件下。大多数现有课程依赖手工设计的启发式方法（如固定的不确定性指标）并遵循静态计划，这无法适应模型不断变化的高维训练动态，导致类别偏差。受强化学习启发，我们将课程学习构建为序列决策问题，并提出一种自主类别调度器。该调度器包含两个组件：(i) 高维状态编码器，将模型训练状态映射到潜在空间并提炼指示进展的关键特征；(ii) 类别公平的策略梯度目标，确保各类别间的平衡提升。结合混合源-目标监督，学习到的类别排序在每个阶段引导网络关注最具信息量的类别，实现更自适应和动态的学习。值得注意的是，我们的方法在三个广泛使用的基准数据集（如ACDC、Dark Zurich和Nighttime Driving）上达到最先进性能，并在合成到真实的语义分割中展现出泛化能力。
URL
https://arxiv.org/abs/2603.24322
