Google｜“自动机器学习”取得重大突破：机器设计的机器学习软件已达AI专家设计水平！

－－－全球人工智能编译－－－

参与：邵明  王健

       Google和其他一些公司都认为软件能够学习，并且能够做一些以前只能由人工智能专家做的事情。

       人工智能的进步使一些人担忧人工智能在不久之后会取代人类的一些工作，如无人驾驶，这将造成数万人失业。现在研究人员发现，人工智能也能够学会处理开发人工智能软件中最棘手的问题：机器学习软件的设计任务。

       在最近的一次实验中，谷歌大脑的人工智能研究组的研究人员让软件自行设计了一个机器学习软件，并让该软件对语言处理进行基准测试。结果表明，其性能不亚于目前由人设计的任何机器学习软件。

      最近几个月，其它一些研究小组也报道了软件学会制作软件的进展。这包括：ElonMusk联合创办的非盈利性研究机构OpenAI，麻省理工学院，加州大学伯克利分校和谷歌的人工智能研究小组DeepMind的研究人员。

       如果这项技术变得成熟，这无疑将加快机器学习软件的开发速度，并对经济的发展带来难得的机遇。人工智能软件公司应该加大对机器学习领域人才的重视。

   谷歌大脑研究小组的首席科学家Jeff  Dean上周表示，软件学会制作软件（称为自动机器学习）是人工智能领域最有前景的研究任务之一，他带领的团队正在积极探索这一领域的发展。

      Jeff  Dean在加利福尼亚州AI Frontiers会议上说，机器学习的专家、数据和计算能力以前是机器学习三个必不可少的要素，现在“自动机器学习”或许能改变这一切。

       谷歌的DeepMind的一组实验表明，“自动机器学习”能够帮助极大地减小机器学习过程中对数据的依赖，以前通常靠大量的数据训练来提升机器学习的性能这一现状将会有所缓解，这无疑会大机器学习的步伐。

      研究人员挑战传统，使用软件自行设计机器学习系统，能执行了许多不同的任务，如导航迷宫，语言处理。这些实验都体现了这项技术不仅具有很好的泛化能力，并且在执行新任务方面也更简洁（不需要像传统的复杂的额外训练过程）。

      自动机器学习的想法已经有一段时间了，最早产生于20世纪90年代，但是先前的实验结果都不能达到人构建的机器学习系统的水平。加拿大蒙特利尔大学的Yoshua Bengio教授称这真是一项令人兴奋的结果。

      Yoshua Bengio说，深度学习的出现激发了人们对机器学习研究的兴趣，但是这需要强大的计算能力，深度学习对计算能力的极端要求有时使解决某些问题并不现实，实际上自动机器学习还不能代替机器学习专家的工作。从这一角度来说，提高机器学习的水平和开发高性能计算设备显得极端重要。

       谷歌大脑的研究人员曾使用800个高功率图形处理器来驱动机器学习软件，这才构建了能与人类竞争的图像识别系统。

       麻省理工学院媒体实验室的研究员Otkrist Gupta认为这种情况将会有所改变。他和麻省理工学院的其它同事计划对与他们实验有关的软件开源，这些实验包括：机器学习软件自行设计深度学习系统、对象识别等。

        Otkrist Gupta认为企业和研究人员应该找到自动机器学习更加可行的途径，推动机器学习的发展。同时他也认为，减轻数据科学家收集、处理数据的负担将会带来很大的回报，这不仅会让你的模型更加高效，也会让你有更多时间去探索更高层次的想法。

原文来源 technologyreview：

Progress in artificial intelligence causes some people to worry that software will take jobs such as away from humans. Now leading researchers are finding that they can make software that can learn to do one of the trickiest parts of their own jobs—the task of designing machine-learning software.

In one experiment, researchers at the Google Brain artificial intelligence research group had software design a machine-learning system to take a test used to benchmark software that processes language. What it came up with from software designed by humans.

In recent months several other groups have also reported progress on getting learning software to make learning software. They include researchers at the  (which was cofounded by Elon Musk), , the University of California, , and Google’s other artificial intelligence research group, .

If self-starting AI techniques become practical, they could increase the pace at which machine-learning software is implemented across the economy. Companies must currently pay a premium for machine-learning experts, who are in short supply.

Jeff Dean, who leads the Google Brain research group, mused last week that some of the work of such workers could be supplanted by software. He described what he termed “automated machine learning” as one of the most promising research avenues his team was exploring.

“Currently the way you solve problems is you have expertise and data and computation,” said Dean, at the in Santa Clara, California. “Can we eliminate the need for a lot of machine-learning expertise?”

One from Google’s DeepMind group suggests that what researchers are terming “learning to learn” could also help lessen the problem of machine-learning software needing to consume vast amounts of data on a specific task in order to perform it well.

The researchers challenged their software to create learning systems for collections of multiple different, but related, problems, such as navigating mazes. It came up with designs that showed an ability to generalize, and pick up new tasks with less additional training than would be usual.

The idea of creating software that learns to learn has been around for a while, but previous experiments didn’t produce results that rivaled what humans could come up with. “It’s exciting,” says , a professor at the University of Montreal, who previously explored the idea in the 1990s.

Bengio says the more potent computing power now available, and the advent of a technique called , which has sparked recent excitement about AI, are what’s making the approach work. But he notes that so far it requires such extreme computing power that it’s not yet practical to think about lightening the load, or partially replacing, machine-learning experts.

Google Brain’s researchers describe using 800 high-powered graphics processors to power software that came up with designs for image recognition systems that rivaled the best designed by humans.

Otkrist Gupta, a researcher at the MIT Media Lab, believes that will change. He and MIT colleagues plan to open-source the software behind their , in which learning software designed deep-learning systems that matched human-crafted ones on standard tests for object recognition.

Gupta was inspired to work on the project by frustrating hours spent designing and testing machine-learning models. He thinks companies and researchers are well motivated to find ways to make automated machine learning practical.

“Easing the burden on the data scientist is a big payoff,” he says. “It could make you more productive, make you better models, and make you free to explore higher-level ideas.”

全球人工智能近期经典文章推荐