与拥有庞大IT产业的中国等其他国家相比，为什么印度在人工智能领域落后了

Why India is not catching up in field of AI when compared to other countries like China, having a big IT industry?

与拥有庞大IT产业的中国等其他国家相比，为什么印度在人工智能领域落后了？

以下是Quora网友的评论：

Krishna Kumar Subramanian

You’re going to find this shocking.

As the U.S. seeks to contain China’s progress in artificial intelligence through sanctions, OpenAI Chief Executive Sam Altman is choosing engagement.

以下这个事实足以会让你震惊：

美国希望通过制裁遏制中国在人工智能领域的崛起，但与此同时OpenAI首席执行官萨姆·奥特曼却选择跟中国进行接触。

Dialing in from overseas to a packed conference in Beiing on Saturday to widespread cheers in the audience, Altman emphasized the importance of collaboration between American and Chinese researchers to mitigate the risks of AI systems, against a backdrop of escalating competition between Washington and Beiing to lead in the technology.

奥特曼上周六在北京参加了一场会议，台下座无虚席，气氛热烈。他强调了美中研究人员密切合作、降低人工智能系统风险的重要性。目前，美国和中国为争夺该技术的领先地位而展开激烈竞争。

“China has some of the best AI talent in the world,” Altman said. “So I really hope Chinese AI researchers will make great contributions here.”

OpenAI doesn’t make available its services, including ChatGPT, in China.

Altman and Geoff Hinton, a so-called godfather of AI who quit Google to warn of the potential dangers of AI, were among more than a dozen American and British AI executives and senior researchers who spoke at the conference.

“中国拥有一批全球最优秀的人工智能人才，”奥特曼说。“所以我真心希望中国的人工智能研究人员能做出巨大贡献。”

OpenAI在中国尚未提供包括ChatGPT在内的服务。

奥特曼和杰夫·辛顿等十多名美国和英国的人工智能高管和高级研究人员在大会上进行发言。杰夫·辛顿被誉为人工智能教父，他向世界警示人工智能的潜在危险。

Chinese speakers at the conference came from top universities and companies including U.S.-blacklisted telecom company Huawei Technologies, and speech-recognition firm iFlytek.

“This event is extremely rare in U.S.-China AI conversations,” said Jenny iao, a partner at venture-capital firm Leonis Capital who researches AI and China. “It’s important to bring together leading voices in the U.S. and China to avoid issues such as AI arms racing, competition between labs and to help establish international standards,” she added.

与会的中国发言人来自顶尖大学和公司，包括被美国列入黑名单的电信公司华为技术有限公司和语音识别公司科大讯飞。

风投公司Leonis Capital负责研究人工智能和中国的合伙人Jenny X iao表示：“此次大会在美中人工智能对话中十分难得。”她补充说：“美国和中国的领军人物会面交流，此次大会具有重大意义，避免了两国之间人工智能军备竞赛、实验室竞争等问题，有助于打造国际标准。”

By some metrics, China now produces more high-quality research papers in the field than the U.S. but still lags behind in “paradigm-shifting breakthroughs,” according to an analysis from the Brookings Institution

India figures nowhere in this conference.

How long will our leadership go on preening itself at overseas NRI conferences and neglect urgent investments in new technologies?

And what is the Scientific Advisor to the Government of India doing?

根据布鲁金斯学会的一项分析，在某些指标方面，中国现在在人工智能领域发表的高质量研究论文的数量已超过美国，但在“转变突破”方面仍落后于美国。

印度在这次会议中无足轻重。

我们的领导层还要继续自吹自擂多久呢？什么时候才能开始对新技术进行投资？

印度政府的科学顾问又在忙啥呢？

Whitchurch Rajkumar

AI takes a lot of specialized skills and a highly creative analytical brain along with pattern recognition capabilities for a student to master apply. It also needs a generalist who can seamlessly move between Robotics , ML and RL.

These are skills that require India to create a dedicated college degree focusing on AI+CS+mech or AI+EE etc.

人工智能需要大量的专业技能和高度创新的分析人才，以及模式识别能力，同时还需要能够在机器人、ML和RL之间无缝转变的通才。

这就要求印度必须创建一个专门教授AI+CS+mech或AI+EE等技术的大学学位。

IT work is not that complicated there learning curve And the mathematical foundation is too basic.

As an MS AI/Robotics graduate I can say the India is facing an uphill task. Only way out is for local students with US degrees like myself willing to start something in India.

We need to start like China Initially as copycat companies using pre-exsting techniques to build AI solutions and robots build an ecosystem before the innovation can spring forth.

IT工作并没有那么复杂，学习曲线和数学基础太简单了。

作为一名人工智能/机器人专业的毕业生，我可以说印度正面临着一项艰巨的任务。唯一的希望是像我这样拥有美国学位的印度学生愿意在印度开始努力。

我们必须像中国一样从头开始，先模仿学习，使用现有的技术来构建人工智能解决方案和机器人，在实现创新前搭建好生态系统。

Finally you can’t train college grads for 6 months and expect to build AI products. IT companies like TCS infy used to Have 6 month training on apis or some Java/C# before deploying new hires into projects. Unfortunately that does not work in AI /ML , sure you can train a new hire on tensorflow api but using it requires deep theoretical mathematics and high degree of comfort with linear algebra, calculus etc which is not easy to teach of the student is clueless about it.

最后，你也不可能为大学毕业生速成培训6个月后就期望他们能成功构建人工智能产品。像TCS infy这样的IT公司都会在将新员工安排到具体项目之前进行6个月的api或Java/ c#培训。但很可惜，这在AI /ML中没什么用，当然了，你可以培训新员工使用tensorflow api，但这项技术需要深厚的理论数学和对线性代数、微积分等的高度熟悉，难度很大。

Ben Podgursky

Which country is leading in AI research?

China (by one count). The US (by the other count)

哪个国家在AI领域处于领先地位？

Alibaba (China)

Amazon (US)

Apple (US)

Baidu (China)

Facebook (US)

Google (US)

IBM (US)

Microsoft (US)

Tencent (China)

阿里巴巴(中国)

亚马逊(美国)

苹果(美国)

百度(中国)

Facebook(美国)

谷歌(美国)

IBM(美国)

微软(美国)

腾讯(中国)

These are the nine companies that matter. (Source: Nine Companies Are Sha The Future Of Artificial Intelligence). Neither the US nor Chinese tech giants have any competition.

But classifying AI research by “country” isn’t really correct. Neither the Chinese nor US governments have any AI sophistication. 90% of the sophistication exsts in these nine companies; the remainder lives in academic institutions, where the US (probably) still has a small lead.

以上九家都是具有重要地位的公司。(资料来源：九家公司正在塑造人工智能的未来)。美国和中国的科技巨头们没有其他竞争对手。

但将人工智能研究按”国家”分类并不一定正确。中国和美国政府都不具备任何人工智能技术。90%的技术都是由以上九家公司掌握，剩下10%的技术则是由学术机构掌握，在这一方面美国(很可能)仍略微领先。

The real question is “to what degree does corporate AI sophistication impart an advantage to the government in question.” Historically, strong companies are available as a resource to the government; during the Cold War, a strong Boeing meant a strong US.

In China, this certainly remains true. Chinese conglomerates act as an arm of the government. Baidu’s AI talent 100% translates into strength for the PRC military.

真正的问题在于”企业人工智能的复杂性在多大程度上给政府带来了优势。”历史上，大企业可以作为政府的资源；在冷战时期，强大的波音公司就代表着强大的美国。

在中国，这当然也是事实。中国的综合性大企业也被视为政府的左膀右臂。AI技术得以100%转化为中国军队的实力。

In the US, this is increasingly not true. Google, the leader among the companies in question, has completely pulled out of defense contracts, out of concern by engineers that their work will be used for military purposes. This isn’t isolated to Google.US engineers have the leverage to force corporate decisions in a way completely alien to Chinese engineers.

但在美国，这一点越来越难实现了。在这些公司中处于领先地位的谷歌公司已经完全退出了国防合作，工程师们担心他们的工作成果会被用于军事目的。这一点不单单发生在谷歌身上。美国工程师有能力以一种中国工程师闻所未闻的方式迫使企业做出决定。

Long story short:

US tech companies still hold a lead in the actual technology. Google has pulled ahead with DeepMind and TPUs. Chinese tech companies have not caught up (yet).

China, as a government, has access to this technology in a way the US government does not.

长话短说：

美国科技公司在技术运用方面仍处于领先地位。谷歌已经领先于DeepMind和TPUs。中国科技公司还没能赶上。

但中国的政府可以用美国政府做不到的方式获取这种技术。

This may evolve if the US puts more pressure on tech companies to license their technology to ends the engineers do not want.

如果美国对科技公司施加更大压力，要求它们将技术授权给工程师们反对的对象，这种情况可能会发生变化。

Rajarshi

How is AI research in India different from AI research in US or China?

印度的人工智能研究与美国或中国的人工智能研究有什么不同？

It's different because it's lagging. A lot.

India is, what you could call, in a “Mad rush to deployment of new technology.” 58 percent of Indian companies are using AI for commercial use, while Australia is at 49 percent, and US just on 32.

不同之处就在于印度落后了，落后很多。

你可以称之为，印度正处于“新技术应用的疯狂热潮”之中。”58%的印度公司都把人工智能用于商业用途，这个比例在澳大利亚为49%，而美国仅为32%。

Thing is, we don't know how to make this stuff. Little of whatever ks being used is actually made in India. Mostly it comes from MNCs like Accenture, Microsoft, etc. While these guys employ Indian scientists in India under their huge R&D facilities, the country's own AI investment is just 180 million dollars, as opposed to the 1.62 billion dollars across the World.

但问题是，我们不知道怎么制造这些东西。几乎找不到印度制造的同类产品，主要都是埃森哲、微软等跨国公司的产品。虽然这些企业设在印度的大型研发中心也雇佣了不少印度科学家，但印度自己的人工智能投资规模只有1.8亿美元，其他国家的投资已达16.2亿美元。

So, why can't we make AI? Well, to understand that, you need to know, theoretically how AI works. Deep learning is a technique of self-correction mechanism. It learns stuff based on the previous, collected instances. An AI will take a decision in future based on whatever instances of past decision making has been fed to it.

那么，为什么我们造不出人工智能的产品呢？要理解这一点，你必须知道，理论上人工智能是如何工作的。深度学习是一种有关自我修正机制的技术，它会根据前期收集到的实例学习信息。人工智能会根据过去的实例经验做出决定。

So, naturally, the more instances, or data you provide, the more accurately your AI is going to perform. (I say your AI because most neural networks are actually open ended, which means it's architecture or code is editable, so you could tweak it as per your wish.)

This was a great barrier for furnishing this data. But then something amazing happened. In the West. Social media exploded.

所以，你提供的实例或数据越多，AI就能执行得越准确。(因为大多数类脑计算模型都是开放的，这意味着类脑计算模型的架构或代码都是可编辑的，所以你可以根据自己的想法进行调整。)

这本来是提供数据的一大障碍。但神奇的事情发生了，社交媒体在西方国家突然流行开来。

Yeah. Facebook. It entered the market, and suddenly the internet was a plethora of data, or what we call “Big data.” Every single thing you upload on facebook, right from checking into some flimsy restaurant to posting useless relationship stuff, helps the neural networks to gain some “insights” about the working of human mind.

(I agree that the AI doesn't seem to be learning from the “Greatest minds”, I too hope they switch the data collection platform from facebook to quora. lol)

没错，就是Facebook。它进入市场后，互联网上突然出现了大量的数据，我们称之为“大数据”。你在Facebook上发布的所有内容，从打卡某家餐厅到发布无聊的人际关系信息，都能让类脑计算模型神经网络学习到关于人类思维活动的”内幕”。

(我认为人工智能似乎没有向”最伟大的头脑”学习，我也希望他们能把数据收集平台从Facebook切换到quora)

Anyways,poor India lagged behind because of this very fact. We JUST don't have enough raw data! We don't even have a search engine. US has facebook, google (see how Google fethes the most relavant links? That's AI), China has Ali Baba and Baidu. (Incidentally, KFC has teamed up with Baidu, and they're going to develop a system by which, whenever you're in front of the counter, the face recognition ks going to identify you, recall your last orders, sense your mood and suggest items relevant to it. What!?!)

总之，印度的落后就是因为这个事实。我们没有足够的原始数据！因为我们连搜索引擎都没有。美国有Facebook，谷歌，中国有阿里巴巴和百度。(顺便说一句，肯德基与百度合作，他们计划开发一个系统，有了该系统，无论你何时出现在柜台前，人脸识别系统都能识别你的身份，调取你最近一次订单的信息，感知你的情绪，并推荐合适的商品。什么! ? !)

So, that's that. India, having world's biggest android population is generating the data that is restricted for access to Indian industries.

I don't know if it's a blessing in disguise, seeing how AI looks like it's going to kill the job industry forever, but that's another day's story.

综上，印度拥有世界上最大的安卓用户规模，不断生成阻碍印度进入人工智能行业的数据。

我不知道这是否算得上因祸得福，因为人工智能似乎会抹杀就业，不过这已经是另一回事了。