arXiv 2603.235732026-03-24cs.LGcs.AI

双标准课程学习：应用于时间数据

Dual-Criterion Curriculum Learning: Application to Temporal Data

Université Paris Saclay, Université Paris Cité, ENS Paris Saclay, CNRS, SSA, INSERM, Centre Borelli（巴黎萨克雷大学、巴黎cité大学、巴黎萨克雷高等师范学院、国家科学研究中心、SSA、国家医学研究院、Borelli中心）；Centre d’Analyse et de Mathématique Sociales, EHESS, CNRS（分析与数学社会研究中心、高等社会科学学院、国家科学研究中心）；EDF R&D（电力公司研发）

Gaspard Abel, Eloi Campagne, Mohamed Benloughmari, Argyris Kalogeratos

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出双标准课程学习框架，结合损失和密度标准提升时间序列预测的训练效果。

摘要：

课程学习（CL）是一种元学习范式，通过按难度顺序逐步训练模型来提升性能。定义有意义的难度评估措施至关重要，通常是有效学习的主要瓶颈，而常用启发法往往仅适用于特定应用。本文提出双标准课程学习（DCCL）框架，结合实例难度的两种评估视角：损失标准与在数据表示空间中学习的密度标准。本质上，DCCL在考虑数据稀疏性放大学习难度的情况下校准训练证据（损失）。作为测试平台，我们选择时间序列预测任务。我们在标准的单次通过和婴儿步骤训练计划下评估了我们的框架。实验证明，在这种设置下，基于密度和混合双标准课程优于仅损失的基线和标准非CL训练。

英文摘要：

Curriculum Learning (CL) is a meta-learning paradigm that trains a model by feeding the data instances incrementally according to a schedule, which is based on difficulty progression. Defining meaningful difficulty assessment measures is crucial and most usually the main bottleneck for effective learning, while also in many cases the employed heuristics are only application-specific. In this work, we propose the Dual-Criterion Curriculum Learning (DCCL) framework that combines two views of assessing instance-wise difficulty: a loss-based criterion is complemented by a density-based criterion learned in the data representation space. Essentially, DCCL calibrates training-based evidence (loss) under the consideration that data sparseness amplifies the learning difficulty. As a testbed, we choose the time-series forecasting task. We evaluate our framework on multivariate time-series benchmarks under standard One-Pass and Baby-Steps training schedules. Empirical results show the interest of density-based and hybrid dual-criterion curricula over loss-only baselines and standard non-CL training in this setting.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。