arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1903.01092cs.CV
收藏

Zero-Shot任务转移

Zero-Shot Task Transfer

Indian Institute of Technology, Hyderabad(印度理工学院海得拉巴分校)

Arghya Pal, Vineeth N Balasubramanian

摘要:

在这项工作中,我们提出了一种新的Meta学习算法,即TTNet,回归模型参数的新任务,没有地面真理是可用的(zero-shot任务)。为了适应新的zero-shot任务,我们的元学习器从已知任务的模型参数(具有地面真实值)和已知任务与zero-shot任务的相关性中学习。这种直觉在认知科学中找到了立足点,在认知科学中,主体(人类婴儿)可以通过将其与旧概念(手部运动或自我运动)相关联来适应新概念(深度理解),而无需接受明确的监督。我们在Taskonomy数据集上评估了我们的模型,其中有四个任务为zero-shot:表面法线,房间布局,深度和相机姿态估计。这些任务是基于数据采集复杂性和与使用深度网络的学习过程相关联的复杂性来选择的。我们提出的方法在我们的每个zero-shot任务上的表现优于最先进的模型(使用地面实况),显示了对zero-shot任务传输的承诺。我们还进行了大量的实验,研究我们的方法的各种选择,以及如何提出的方法也可以用于迁移学习。据我们所知,这是第一次在任务空间中进行零样本学习。

英文摘要:

In this work, we present a novel meta-learning algorithm, i.e. TTNet, that regresses model parameters for novel tasks for which no ground truth is available (zero-shot tasks). In order to adapt to novel zero-shot tasks, our meta-learner learns from the model parameters of known tasks (with ground truth) and the correlation of known tasks to zero-shot tasks. Such intuition finds its foothold in cognitive science, where a subject (human baby) can adapt to a novel-concept (depth understanding) by correlating it with old concepts (hand movement or self-motion), without receiving explicit supervision. We evaluated our model on the Taskonomy dataset, with four tasks as zero-shot: surface-normal, room layout, depth, and camera pose estimation. These tasks were chosen based on the data acquisition complexity and the complexity associated with the learning process using a deep network. Our proposed methodology out-performs state-of-the-art models (which use ground truth)on each of our zero-shot tasks, showing promise on zero-shot task transfer. We also conducted extensive experiments to study the various choices of our methodology, as well as showed how the proposed method can also be used in transfer learning. To the best of our knowledge, this is the firstsuch effort on zero-shot learning in the task space.

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。

↑