Active Learning by Acquiring Contrastive Examples
Katerina Margatina† Giorgos Vernikos‡∗ Loïc Barrault† Nikolaos Aletras† †University of Sheffield ‡EPFL ∗HEIG-VD {k.margatina, l.barrault,
Abstract
Common acquisition functions for active learning use either uncertainty or diversity sampling, aiming to select difficult and diverse data points from the pool of unlabeled data, respectively. In this work, leveraging the best of both worlds, we propose an acquisition function that opts for selecting contrastive examples, i.e. data points that are similar in the model feature space and yet the model outputs maximally different predictive likelihoods. We compare our approach, Cal (Contrastive Active Learning), with a diverse set of acquisition functions in four natural language understanding tasks and seven datasets. Our experiments show that Cal performs consistently better or equal than the best performing baseline across all tasks, on both in-domain and out-of-domain data. We also conduct an extensive ablation study of our method and we further analyze all actively acquired datasets showing that Cal achieves a better trade-off between uncertainty and diversity compared to other strategies.
中文速览
主动学习（active learning）领域长期存在一个两难困境：基于不确定性的采样容易反复挑选相似的难例，而基于多样性的采样又容易选出对模型提升无益的简单样本。这篇论文提出了一种名为 CAL（Contrastive Active Learning，对比主动学习）的新采集函数，其核心思路是：从未标注数据池中挑选那些在模型特征空间里与已标注样本相似、但模型给出的预测概率分布却截然不同的"对比样本"——这类样本很可能处于模型决策边界附近，信息量最大。具体做法是用 BERT 的 [CLS] 向量为每个未标注候选样本找到 K 个最近邻的已标注样本，再用 KL 散度衡量两者预测分布的差异，得分最高的样本优先被选中标注。在情感分析、主题分类、自然语言推理和复述检测共七个数据集上的实验表明，CAL 在域内和域外评估中均持续优于或持平于所有对比基线，同时在不确定性与多样性之间取得了更好的平衡，为低标注预算场景下的自然语言理解任务提供了一种实用且高效的数据采集策略。
原文 arXiv:2109.03764;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2109.03764v1