arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1707.03017cs.CVcs.AIcs.CLstat.ML
收藏

没有强先验知识的视觉推理学习

Learning Visual Reasoning Without Strong Priors

MILA(MILA蒙特利尔学习算法研究所);Rice University(莱斯大学);Univ. Lille(里尔大学);CNRS(法国国家科学研究中心);Centrale Lille(里尔中央理工学院);Inria(法国国家信息与自动化研究所);CIFAR(加拿大高等研究院)

Ethan Perez, Harm de Vries, Florian Strub, Vincent Dumoulin, Aaron Courville

摘要:

实现人工视觉推理-回答需要多步骤,高层次过程的图像相关问题的能力-是迈向人工通用智能的重要一步。这种多模态任务需要学习一个依赖于问题的,结构化的推理过程,从语言的图像。标准的深度学习方法倾向于利用数据中的偏差,而不是学习这种底层结构,而领先的方法学习成功地进行视觉推理,但却是手工制作的推理。我们表明,一个通用的,有条件的批量归一化方法实现了最先进的结果,CLEVR视觉推理基准与2.4%的出错率。我们的表现优于下一个最好的端到端方法(4.5%),甚至是使用额外监督的方法(3.1%)。我们探索了我们的模型,以揭示它是如何推理的,表明它已经学会了一个依赖于问题的多步骤过程。以前的工作已经运作的假设下,视觉推理要求一个专门的架构,但我们表明,一个一般的架构与适当的条件可以学习视觉上的原因有效。

英文摘要:

Achieving artificial visual reasoning - the ability to answer image-related questions which require a multi-step, high-level process - is an important step towards artificial general intelligence. This multi-modal task requires learning a question-dependent, structured reasoning process over images from language. Standard deep learning approaches tend to exploit biases in the data rather than learn this underlying structure, while leading methods learn to visually reason successfully but are hand-crafted for reasoning. We show that a general-purpose, Conditional Batch Normalization approach achieves state-of-the-art results on the CLEVR Visual Reasoning benchmark with a 2.4% error rate. We outperform the next best end-to-end method (4.5%) and even methods that use extra supervision (3.1%). We probe our model to shed light on how it reasons, showing it has learned a question-dependent, multi-step process. Previous work has operated under the assumption that visual reasoning calls for a specialized architecture, but we show that a general architecture with proper conditioning can learn to visually reason effectively.

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。

↑