阅读清单 2017-06-10

「阅读清单」将不定期筛选全球最具阅读价值的 3 到 5 篇科技深度文章。今天你将看到如下三篇文章:

  • 苹果在人工智能赛道的三年狂奔;
  • GIF 三十岁生日快乐;
  • 真正人脑级别的人工智能何时出现?

苹果在人工智能赛道的三年狂奔

当 2010 年 Siri 问世时,《哈佛商业评论》高呼: Siri is as revolutionary as the Mac。但随后 Siri 就因为太傻而被世人嘲笑,直到 2014 年,苹果公司开始重构 Siri 的底层架构,也是从这一年开始,苹果的人工智能正式起步,而此时,它显然已经掉队。

这篇来自 engadet 的报道汇总了这三年时间里,苹果公司如何快速追赶,其中 2014 年重构 Siri 是一个里程碑的驾驶,去年 Steven Levy 曾专门写过一篇长文介绍这个项目

自此开始,苹果的追赶速度开始加快,尽管直到 2015 年冬天,外界还极度不看好苹果的人工智能,比如多伦多大学计算机教授 Richard Zemel 这样说:

Apple is off the scale in terms of secrecy,They‘re completely out of the loop.

另一位深度学习领域大拿 Yoshua Bengio 也说:

There’s no way they can just observe and not be part of the community and take advantage of what is going on,I believe if they don‘t change their attitude, they will stay behind.

但到了 2016 年,苹果的步伐显然加快了,比如去年 WWDC 上,苹果公司介绍了一种全新的「大数据」—「differential privacy」,这个词翻译成人话就是在基于统计学原理,在一个群体层面收集数据,而非像FB、Google 那样收集个体的数据。你可以理解为是苹果重新定义了「大数据」。至于好处呢,Wired 这样说:

…But it can’t extract anything about a single, specific one of those people that might represent a privacy violation. And neither, in theory, could hackers or intelligence agencies.

这是苹果对于自身坚持的「设备智能」的新方法,随后苹果又有两大举措:

After buying Seattle-based machine learning AI startup Turi for around $200 million in August 2016, Apple hired away from Carnegie Mellon University that October.

到了 2016 年年度深度学习顶级峰会 Neural Information Processing Systems 上,Russ Salakhutdinov 表示苹果公司即将发布人工智能研究论文。所有这些都在 2017 年的 WWDC 上有了充分展示,而传闻中的专属人工智能的芯片还未曝光,是会用到新一代 iPhone 上吗?

GIF 三十岁生日快乐

GIF 的历史居然已经三十年了,早在还没有互联网的 1987 年:

Developer Steve Wilhite and his team at tech giant CompuServe had a problem to solve: how to make a computer display an image while also saving memory. It was 1987, four years before the advent of the World Wide Web, when users who wanted to access email or transfer files did so with hourly subscriptions from companies like CompuServe. Then as now, the issue was space. How could a color image file be shared without taking up too much of the computer’s memory? Wilhite found a way to do so using a compression algorithm (more on this soon) combined with image parameters like the number of available colors (256). His new creation could be used for exchange images between computers, and he called it Graphics Interchange Format. The GIF was born.”

人们对 GIF 的态度也是过去三十年「内容」产业的缩影:

“Before, GIFs were dressing up the content,” says Jason Eppink, curator of digital media at the Museum of Moving Images. GIFs were clip-art images and construction symbols, he explains. But now—“the GIF itself has become the destination.”

真正人脑级别的人工智能何时出现?

目前公众对于人工智能的认知绝大多数停留在好莱坞大片的层面,而在包括霍金、马斯克等「外行人」眼里1,人工智能的邪恶一面似乎触手可及,但在人工智能领域的专家眼里,第一个门槛就是什么时候能出现人脑级别的人工智能?

IEEE spectrum 为此专访了多位该领域的专家,答案并不一致,这篇文章的价值并不是提供一个准确的答案(事实上,也不会有准确答案),而是提出了两个颇有意义的思考:

  • 其一,人工智能一定和人脑智能一样吗?
  • 其二,假如按照最乐观的预计,库兹韦尔认为会在 2029 年,那么接下来的 12 年,我们到底该怎么看待机器?

  1. 我曾专门分析过霍金在人工智能上的错误观点,详见第 68 期会员通讯。 

Zhao Saipo

View posts by Zhao Saipo
赵赛坡,科技博客作者、资深科技观察家、付费科技评论 Dailyio 创始人、出品人,覆盖 3000+ 付费用户。 曾担任 TechTarget 中国区记者、频道主编、AI 自媒体「机器之心」前联合创始人。