arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.00743cs.CVcs.AIcs.CLcs.LG
收藏

迈向通用视觉系统

Towards General Purpose Vision Systems

Allen Institute for AI(艾伦人工智能研究所);University of Illinois at Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校)

Tanmay Gupta, Amita Kamath, Aniruddha Kembhavi, Derek Hoiem

摘要:

今天的计算机视觉系统主要是N目的系统,为预定义的一组任务而设计和训练。使这样的系统适应新的任务是具有挑战性的,并且通常需要对网络架构(例如添加新的输出头)或训练过程(例如添加新的损耗)进行非平凡的修改。为了减少开发新应用程序所需的时间和专业知识,我们希望创建通用视觉系统,该系统可以学习和执行一系列任务,而无需对架构或学习过程进行任何修改。在本文中,我们提出了GPV-1,一个任务无关的视觉语言架构,可以学习和执行任务,涉及接收图像和产生文本和/或边界框,包括分类,定位,视觉问答,字幕,等等。我们还提出了评估的一般性的架构,技能概念转移,学习效率,可能会通知未来的工作,通用视觉。实验结果表明,GPV-1在多个任务上都是有效的,跨任务重用了一些概念知识,可以执行引用表达式任务zero-shot,并且在使用少量训练样本的情况下进一步提高了zero-shot性能。

英文摘要:

Computer vision systems today are primarily N-purpose systems, designed and trained for a predefined set of tasks. Adapting such systems to new tasks is challenging and often requires non-trivial modifications to the network architecture (e.g. adding new output heads) or training process (e.g. adding new losses). To reduce the time and expertise required to develop new applications, we would like to create general purpose vision systems that can learn and perform a range of tasks without any modification to the architecture or learning process. In this paper, we propose GPV-1, a task-agnostic vision-language architecture that can learn and perform tasks that involve receiving an image and producing text and/or bounding boxes, including classification, localization, visual question answering, captioning, and more. We also propose evaluations of generality of architecture, skill-concept transfer, and learning efficiency that may inform future work on general purpose vision. Our experiments indicate GPV-1 is effective at multiple tasks, reuses some concept knowledge across tasks, can perform the Referring Expressions task zero-shot, and further improves upon the zero-shot performance using a few training samples.

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。

↑