EMNLP Findings2020
Dynamic Data Selection for Curriculum Learning via Ability Estimation
Abstract
Curriculum learning methods typically rely on heuristics to estimate the difficulty of training examples or the ability of the model. In this work, we propose replacing difficulty heuristics with learned difficulty parameters. We also propose Dynamic Data selection for Curriculum Learning via Ability Estimation (DDa-CLAE), a strategy that probes model ability at each training epoch to select the best training examples at that point. We show that models using learned difficulty and/or ability outperform heuristic-based curriculum learning models on the GLUE classification tasks.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
6 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
Cited by top-tier papers6
- Training Dynamics for Curriculum Learning: A Study on Monolingual and Cross-lingual NLUFenia Christopoulou, Gerasimos Lampouras, Ignacio IacobacciEMNLP 2022 · 3 citations
- Learning to Solve Complex Problems via Dataset DecompositionWanru Zhao, Lucas Page-Caccia, Zhengyan Shi, Minseon Kim et al.NeurIPS 2025 · 3 citations
- Curriculum Learning for Graph Neural Networks: A Multiview Competence-based ApproachNidhi Vakil, Hadi AmiriACL 2023 · 2 citations
- HuCurl: Human-induced Curriculum DiscoveryMohamed Elgaar, Hadi AmiriACL 2023 · 1 citation
- Ling-CL: Understanding NLP Models through Linguistic CurriculaMohamed Elgaar, Hadi AmiriEMNLP 2023 · 1 citation
Related papers
- Uncertainty-Aware Curriculum Learning for Neural Machine TranslationYikai Zhou, Baosong Yang, Derek F. Wong, Yu Wan et al.ACL 2020 · 78 citations
- Adaptive Curriculum LearningYajing Kong, Liu Liu, Jun Wang, Dacheng TaoICCV 2021 · 61 citations
- Curriculum Learning for Natural Language UnderstandingBenfeng Xu, Licheng Zhang, Zhendong Mao, Quan Wang et al.ACL 2020 · 156 citations
- Curriculum Learning by Dynamic Instance HardnessTianyi Zhou, Shengjie Wang, Jeff A. BilmesNeurIPS 2020 · 113 citations
- SuperLoss: A Generic Loss for Robust Curriculum LearningThibault Castells, Philippe Weinzaepfel, Jérôme RevaudNeurIPS 2020 · 96 citations
