论文交流 >  机器视觉

Label-Free Supervision of Neural Networks with Physics and Domain Knowledge

基于物理和领域知识的无标签监督网络

1年前 904 1  点赞 (0)  收藏 (0)

研究领域: 机器视觉   AAAI2017

应用方向:

原理方法:

软件实现:

论文摘要:

In many machine learning applications, labeled data is scarce and obtaining more labels is expensive. We introduce a new approach to supervising neural networks by specifying constraints that should hold over the output space, rather than direct examples of input-output pairs. These constraints are derived from prior domain knowledge, e.g., from known laws of physics. We demonstrate the effectiveness of this approach on real world and simulated computer vision tasks. We are able to train a convolutional neural network to detect and track objects without any labeled examples. Our approach can significantly reduce the need for labeled training data, but introduces new challenges for encoding prior knowledge into appropriate loss functions.

论文精要:

论文点评 

您可以在评论中对论文进行“摘要翻译” “标签备注” “精要点评” “疑难提问”,我们会及时更新到数据库中!

任何论文都是 "在特定领域内"、"基于某种学术原理"、"研究某个应用问题", 因此分 领域标签 / 应用标签 / 原理标签 / 补充标签

Antares 10个月前

在许多机器学习应用中,带标签的数据数量稀少,而想要获得更多的标签需要付出高昂的代价。我们引入了一种新的神经网络监督学习方法,不同于采用直接给出输入-输出对的直接示例的传统方案。这种方法通过特定的约束条件来指定输出空间,而这些约束条件来源于先前的特定领域知识,例如已知的物理定律。我们展示了这种方法在现实世界和模拟计算机视觉任务的有效性。利用这种方法,我们可以在训练样本没有带任何标签的情况下,成功训练了一个卷积神经网络来检测和跟踪对象。这一方法还可以显着减少对标记的训练数据的需要,并同时带来了将先验知识编码为适当的损失函数的新挑战。


(0) 回复