IEEE Trans Image Process. 2021;30:8332-8341. doi: 10.1109/TIP.2021.3114987. Epub 2021 Oct 5.
Raven's Progressive Matrices (RPM) is highly correlated with human intelligence, and it has been widely used to measure the abstract reasoning ability of humans. In this paper, to study the abstract reasoning capability of deep neural networks, we propose the first unsupervised learning method for solving RPM problems. Since the ground truth labels are not allowed, we design a pseudo target based on the prior constraints of the RPM formulation to approximate the ground-truth label, which effectively converts the unsupervised learning strategy into a supervised one. However, the correct answer is wrongly labelled by the pseudo target, and thus the noisy contrast will lead to inaccurate model training. To alleviate this issue, we propose to improve the model performance with negative answers. Moreover, we develop a decentralization method to adapt the feature representation to different RPM problems. Extensive experiments on three datasets demonstrate that our method even outperforms some of the supervised approaches. Our code is available at https://github.com/visiontao/ncd.
瑞文标准推理测验（Raven's Progressive Matrices，RPM）与人类智力高度相关，被广泛用于衡量人类的抽象推理能力。在本文中，为了研究深度神经网络的抽象推理能力，我们提出了第一个用于解决 RPM 问题的无监督学习方法。由于不允许使用真实标签，我们设计了一个基于 RPM 公式的先验约束的伪目标来近似真实标签，这有效地将无监督学习策略转化为有监督学习策略。然而，伪目标错误地标记了正确答案，因此噪声对比会导致模型训练不准确。为了解决这个问题，我们提出使用负例来提高模型性能。此外，我们还开发了一种去中心化方法，使特征表示适应不同的 RPM 问题。在三个数据集上的广泛实验表明，我们的方法甚至优于一些有监督方法。我们的代码可在 https://github.com/visiontao/ncd 上获得。
