vue路由匹配实现包容性
通用情报 (General Intelligence)
Welcome to General Intelligence, OneZero’s weekly dive into the A.I. news and research that matters.
欢迎来到 General Intelligence ， OneZero 每周都会深入探讨重要的AI新闻和研究。
The Neural Information Processing Systems (NeurIPS) conference was once a sleepy gathering for well-to-do academics, often held in ski towns to accommodate leisure time before lectures. Over the last eight years, that’s changed. A dramatic rise in the popularity of artificial intelligence has suddenly made NeurIPS one of the hottest conference tickets in the world.
吨他神经信息处理系统(NeurIPS)会议曾经是一个沉睡的聚会为小康做学者，在滑雪镇经常举行，以适应讲座之前的闲暇时间 。 在过去的八年中，情况发生了变化。 人工智能的迅猛发展使NeurIPS突然成为世界上最热门的会议门票之一。
But 2020 is the conference’s most drastic shift yet. Organizers announced this week that NeurIPS, now the world’s best-known A.I. conference, would be completely virtual this year, and with considerably cheaper tickets.
但2020是会议的M O对ST剧烈的转变还。 组织者本周宣布，如今已成为世界上最著名的AI会议的NeurIPS将在今年完全虚拟化 ，并且票价要便宜得多。
The move, borne of Covid-19 fears, might unintentionally be a blueprint for a more inclusive A.I. conference that many in the industry have demanded for years.
此举由Covid-19的担忧所引起，可能无意间成为了行业内许多人要求多年的更具包容性的AI会议的蓝图。
NeurIPS became a pillar of the world’s conversation on cutting edge A.I. research following the A.I. boom that started in 2012, after a team of Canadian researchers proved a fringe technique called deep learning outperformed many other forms of A.I. Tech giants began flocking to the formerly academic event, spending millions to sponsor the conference and recruit the next wave of star PhDs.
自2012年开始掀起 AI繁荣之后，NeurIPS成为世界范围内最前沿AI研究对话的Struts，此前加拿大的一组研究人员证明了一种称为“深度学习”的边缘技术胜过许多其他形式的AI技术巨头开始蜂拥而至。 ，花费数百万美元赞助会议并招募下一波明星博士学位。
In 2017, Intel hired Flo Rida to perform a private concert at its NeurIPS party, while Facebook’s party at the conference featured an “unparalleled exclusive view of the private jet runway at the award-winning Long Beach Airport,” according to the invite.
受邀者称，2017年，英特尔聘请Flo Rida在其NeurIPS派对上进行私人音乐会，而Facebook的聚会在会议上展示了“屡获殊荣的长滩机场私人飞机跑道的无与伦比的独家风景”。
This influx of attention meant NeurIPS has also had to reckon with its shortcomings. The field of A.I. is predominantly male and white, which was glaringly obvious when the industry’s chosen playmakers were all gathered in one place. The conference was also named NIPS until 2018, when the “controversial acronym” was changed to NeurIPS after years of pleas from attendees. The exorbitant cost of conference travel and lodging, as well as visa denials, meant many researchers from around the world simply couldn’t attend, furthering the industry’s homogeneity.
这种关注的涌入意味着NeurIPS也不得不考虑其缺点。 人工智能领域主要是男性和白人，当行业内选定的组织者都聚集在一个地方时，这一点就显而易见。 直到2018年，该会议也被命名为NIPS，在经过多年的与会者恳求之后，“ 有争议的首字母缩写 ”更改为NeurIPS。 会议旅行和住宿费用高昂，以及签证被拒签 ，这意味着来自世界各地的许多研究人员根本无法参加，进一步促进了该行业的同质化。
In response, organizations like Black In AI, LatinX in AI, and Women in ML have created spaces for groups that are underrepresented at NeurIPS, holding their own workshops centering women and people of color, as well as awarding travel grants.
作为响应，像Black In AI ， AI的 LatinX和ML的Women这样的组织为NeurIPS中代表性不足的团体创造了空间，他们举办了以妇女和有色人种为中心的研讨会，并提供了旅行补助金。
Facebook’s party at the conference featured an “unparalleled exclusive view of the private jet runway at the award-winning Long Beach Airport.”
Facebook在会议上的聚会展示了“屡获殊荣的长滩机场私人飞机跑道的无与伦比的独家风景”。
Africa’s largest A.I. conference, Deep Learning Indaba, was also created in part to build a community for those who couldn’t attend NeurIPS, which has almost always been held in North America.
非洲最大的AI会议深度学习Indaba也被创建，部分目的是为无法参加NeurIPS的人们建立一个社区，该社区几乎总是在北美举行。
Going virtual, in addition to a “significantly reduced” registration fee, will make the conference more accessible for those who haven’t been able to attend before. In addition to removing travel and lodging costs, the conference lifted its attendance cap. That’s a big deal: In 2018, for example, tickets for NeurIPS sold out in fewer than 12 minutes.
虚拟化，除了“大大减少”的注册费外，还将使以前无法参加会议的人可以更方便地参加会议。 除了消除差旅和住宿费用，会议还提高了出席人数上限。 这是一个大问题：例如，在2018年，NeurIPS的门票在不到12分钟的时间内售罄。
Much of NeurIPS’ virtual schedule and operation is still undecided, and organizers are actively seeking input for how to make the conference successful. By letting everyone attend, they might already be heading in the right direction.
NeurIPS的许多虚拟日程表和操作仍未决定，组织者正在积极寻求有关如何使会议成功的投入。 通过让所有人参加，他们可能已经朝着正确的方向前进。
Here’s some of the most interesting research of the past week:
这是过去一周中一些最有趣的研究：
开除：采用全卷积变分自编码器进行分集除噪 (DIVNOISING: Diversity Denoising with Fully Convolutional Variational Autoencoders)
From microscope images to optical character recognition, researchers demonstrated how A.I. could improve at removing digital noise from images. This approach only needs the noisy image, and to have a few prior examples of the specific kind of noise it’s trying to remove.
从显微镜图像到光学字符识别，研究人员展示了AI如何改善图像中的数字噪声。 这种方法只需要嘈杂的图像，并具有一些其要消除的特定类型噪声的先前示例。
我们完成了ImageNet吗？ (Are we done with ImageNet?)
ImageNet was built as the digital reference for the physical world, so A.I. could learn to see. Now, Google and DeepMind ask whether deep learning’s most important dataset is ready for retirement. The answer is a not-very-satisfying “yes and no.” In short, the images are good, but the labels telling the algorithm what’s in the images could be updated.
ImageNet被构建为物理世界的数字参考，因此AI可以学会观察。 现在，Google和DeepMind询问深度学习最重要的数据集是否准备退役。 答案是不是很令人满意的“是，不是”。 简而言之，图像是好的，但是告诉算法的图像中的标签可以更新。
编程语言的无监督翻译 (Unsupervised Translation of Programming Languages)
Remember when multiple U.S. states’ unemployment systems were breaking because there were so few coders that worked in COBOL? Well, new research from Facebook is a step toward automatically translating that code into newer, more compatible languages.
还记得美国多个州的失业系统因为在COBOL中工作的编码员很少而崩溃了吗？ 好吧，Facebook的最新研究朝着将代码自动翻译成更新的，兼容的语言迈出了一步。
vue路由匹配实现包容性
