arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.08898cs.LGcs.AIcs.CL
收藏

带辅助任务的Few-Shot共形预测

Few-shot Conformal Prediction with Auxiliary Tasks

Massachusetts Institute of Technology(麻省理工学院)

Adam Fisch, Tal Schuster, Tommi Jaakkola, Regina Barzilay

摘要:

当目标任务可用于训练的数据有限时,我们开发了一种新的共形预测方法。保形预测识别一小组有希望的输出候选,以代替单个预测,并保证该组包含高概率的正确答案。然而,当训练数据有限时,预测集很容易变得过大。在这项工作中,我们获得了更紧密的预测集,同时保持理想的边际保证铸造共形预测作为一个元学习范式的辅助任务的可交换集合。我们的一致化算法简单、快速,并且与底层模型、学习算法或数据集的选择无关。我们证明了这种方法在自然语言处理,计算机视觉和计算化学药物发现中的一些Few-Shot分类和回归任务的有效性。

英文摘要:

We develop a novel approach to conformal prediction when the target task has limited data available for training. Conformal prediction identifies a small set of promising output candidates in place of a single prediction, with guarantees that the set contains the correct answer with high probability. When training data is limited, however, the predicted set can easily become unusably large. In this work, we obtain substantially tighter prediction sets while maintaining desirable marginal guarantees by casting conformal prediction as a meta-learning paradigm over exchangeable collections of auxiliary tasks. Our conformalization algorithm is simple, fast, and agnostic to the choice of underlying model, learning algorithm, or dataset. We demonstrate the effectiveness of this approach across a number of few-shot classification and regression tasks in natural language processing, computer vision, and computational chemistry for drug discovery.

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。

↑