Qihoo 360’s Founder Zhou Hongyi Regards Artificial Intelligence, Not O2O As The Future

摘要: In the past two years, Chinese Internet companies all jumped to provide O2O services such as takeaway, door-to-door car-washing, online movie ticket-booking, etc., believing that there are huge business potentials out there. Qihoo 360, however, remains calm amid this craze for O2O services, since Qihoo 360 believes the future will be based on big data and artificial intelligence technologies, that smart devices will become more diversified in the future, and that smartphones aren’t the end of the history at all.

(Chinese Version)

When the match between Google's Go-playing computer program AlphaGo and Go grandmaster Lee Sedol of South Korea just began, Zhou Hongyi, founder and CEO of Qihoo 360, predicted that Lee Sedol would win in his WeChat friend circle. However, after AlphaGo ended the historic match of the ancient Chinese board game with Lee Sedol by taking a 4-1 lead with its fourth victory in the final match of the best-of-five series, he revealed in an internal letter to Qihoo 360’s employees that he had changed his mind and stood for artificial intelligence. He even predicted in the internal letter some potential user scenarios and trends of artificial intelligence technologies.

According to him, AlphaGo and Lee Sedol’s match marked the beginning of a new era: artificial intelligence, as did the victory of BM's supercomputer Deep Blue over Garry Kasparov, the reigning world chess champion in a six game chess match in 1997.

For Mr. Zhou, although machines beat the human race in both cases, they have entirely different implications:

  • Deep Blue’s victory brought forward a boom of IT infrastructures, Internet and the spread of computers.
  • This match, however, demonstrates the new model of thinking for computers, how it could contribute to the development of deep learning and the spread of nerve network technologies, how speech and picture recognition technologies are going to be integrated to deep learning technologies and be applied in all kinds of new products. It is only a matter of time when products developed based on artificial intelligence technologies will be spread to the general public.

So what is Qihoo 360 going to do to adapt to such trend? Here’s Mr. Zhou’s answer:

1. Setting IoT as part of its strategic plans;

2. Focusing on big data and artificial intelligence technologies;

3. Dedicated to diversifying smart devices.

Thus, Mr. Zhou revealed that Qihoo 360 would focus on two main areas of artificial intelligence in the future: picture recognition and big data technologies.

It is worth mentioning that while many people are talking about the threat of artificial intelligence technologies, Mr. Zhou don’t think so. For him, artificial intelligence is never simply about playing world chess and ancient Chinese board game, or about to be a threat to human beings as is described in several science and fictional movies. Instead, artificial intelligence is going to be a boon for the entire human race, just like the previous industrial revolutions to us.

In addition, he stated in the internal letter that:

In the past two years, Chinese Internet companies all jumped to provide O2O services such as takeaway, door-to-door car-washing, online movie ticket-booking, etc., believing that there are huge business potentials out there. Qihoo 360, however, remains calm amid this craze for O2O services, since we are well aware that delivering takeaway food should never be the ultimate goal of Qihoo 360, and that selling movie tickets are not Qihoo 360 is good at.

Furthermore, Mr. Zhou revealed that Yan Shuicheng, chief scientist of Qihoo 360 and president of Qihoo 360 Artificial Intelligence Institute, had already set up his team and been doing some related research. This is the first time Mr. Zhou mentioned Mr. Yan Shuicheng in public. Who is he? According to public data:

Mr. Yan graduated from Peking University, and is a Ph.D at UIUC. His research fields vary from computing sight, deep learning to multi-media analysis. Purine, the first open-source GPU deep learning system in the world, was developed exactly by his team.

Besides, he was included in Thomson Reuter’s rankings “The Most Influential Scientists in 2014”, and named “ISI Highly-cited researchers” twice (both in 2014 and in 2015). In September, 2015, he was named as president of Qihoo 360 Artificial Intelligence Institute.

Coincidentally, I had an interview with Professor Yan the other days, and he predicted that AlphaGo would win even then the game wasn’t over yet. What’s more surprising is that he predicted that AlphaGo would beat Lee Sedol at 4-1. Why is he so sure?

“What makes AlphaGo excel is its capability to do deep learning. In other words, it is able to commit trial and error and tell if others’ action is good or not and react accordingly,” said him.

Besides, AlphaGo is able to tell when it should focus more on being accurate, and when on speed, so that it could win the match while not spending too much time.

Speaking of why it’s Google instead of other companies who have developed mature neural network algorithm, Professor Yan pointed out that data, model and computer resources are the three defining factors for any research. Compared to ordinary researchers, Internet giants like Google are at an advantage in all three dimensions.

He also explained why it was too early to expect a Chinese researcher to conduct similar research: most researchers in Internet companies are busy thinking about how they could apply artificial intelligence technologies when developing real products and make it more convenient for users to live and work, while only a few researchers at academic institutes can bury themselves totally in such cause. Since people are focusing more on accuracy instead of speed, the competition among researchers and Internet companies aren’t about intelligence, but rather about data and computer resources.

For Professor Yan, artificial intelligence technologies can be mainly applied in the following four aspects: visual understanding, speech recognition, natural language understanding and big data analysis.

However, he predicted that products developed based on artificial intelligence technologies would be basically divided into two groups in the future:

  • Cloud-based: Users could upload pictures, voice or speech, and big data to the cloud, and would receive feedbacks in a short time. The cloud is like a huge library of knowledge where data could be analyzed.
  • Device-based: Users don’t have to upload all the data to the cloud, since their devices could do some basic analysis. Most analysis will be conducted within the device, while only part of it will be completed in the cloud, so that the device could work even when there is no Internet.

Professor Yan also revealed that his team was working on baby learning recently. According to his explanation, baby learning has to do with two kinds of capabilities: to start learning when given limited number of sample; to gradually accumulate experience and learn things, which might look similar to deep learning, but is slightly different.

While deep learning is about learning from trials and errors, baby learning is about finding the right ones and gradually be able to grasp a piece of knowledge.

At last, Professor Yan pointed out that it was natural that people might worry about a new thing at first, but fundamentally, artificial intelligence technologies, whether applied in the cloud or on local devices, whether developed by Internet giants or tiny startups, are used to help people.

Mr. Yan’s opinion coincides with Mr. Zhou’s appeal near the end of his internal letter:

I believe now you can understand why we are not too much focused on O2O takeaway food and online movie ticket-booking services. Every people at Qihoo 360 should set American startups founders and engineers as their example and develop some really breath-taking and meaningful products one day. You should not be devoted to making more money, but rather to making something novel and solving problems for the entire human race.

The following is the full transcript of Mr. Zhou Hongyi’s internal letter:

I believe every one of you must have been following closely to the match between Google's Go-playing computer program AlphaGo and Go grandmaster Lee Sedol of South Korea recently. As we all know, AlphaGo ended the historic match of the ancient Chinese board game with Lee Sedol by taking a 4-1 lead with its fourth victory in the final match of the best-of-five series. Whether we admit it or not, AlphaGo and Lee Sedol’s match marked the beginning of a new era: artificial intelligence.

Back in May, 1997, IBM's supercomputer Deep Blue also defeated Garry Kasparov, the reigning world chess champion, in a six game chess match. That match brought forward a boom of IT infrastructures, Internet and the spread of computers. This match, however, demonstrates the new model of thinking for computers, how it could contribute to the development of deep learning and the spread of nerve network technologies, how speech and picture recognition technologies are going to be integrated to deep learning technologies and be applied in all kinds of new products. It is only a matter of time when products developed based on artificial intelligence technologies will be spread to the general public.

In the past two years, Chinese Internet companies all jumped to provide O2O services such as takeaway, door-to-door car-washing, online movie ticket-booking, etc., believing that there are huge business potentials out there. Qihoo 360, however, remains calm amid this craze for O2O services, since we believe the future will be based on big data and artificial intelligence technologies, and that smart devices will become more diversified in the future, and smartphones aren’t the end of the history at all.

Based on the above belief, we’ve set our focus on two main areas: picture recognition technologies and big data technologies, have set up special R&D teams both in the US and in China and have acquired a series of related companies. Yan Shuicheng, chief scientist of Qihoo 360 and president of Qihoo 360 Artificial Intelligence Institute, has already set up his team and has been doing some related research.

We’ve done even more things up till now, but I shall not tell you all of them since I’m bound by business confidentiality.

This match between AlphaGo and Lee Sedol provides a great opportunity for the general public to get to know artificial intelligence, but we should be aware that artificial intelligence is never simply about playing world chess and ancient Chinese board game, or about to be a threat to human beings as is described in several science and fictional movies. Instead, artificial intelligence is going to be a boon for the entire human race, just like the previous industrial revolutions to us.

During the past two years, we’ve already released a series of smart hardware products, but many of you might treat them merely as tiny and meaningless. However, I don’t think so. For example, Qihoo 360 SmartCamera might be a very tiny product, but it is developed exactly by artificial intelligence technologies, such as picture recognition and big data. As a matter of fact, these technologies are still being rapidly improved every day.

This tiny SmartCamero can detect movable objects, and warn the police if the door or windows are opened when the hosts are out. In the near future, it will be able to differentiate between the voice and face of hosts and strangers, between the smile and tears of kids and detect if a kid or an aged person fall over. In addition, if the SmartCamera can be equipped with wheels, it will even be able to go around the house to see if there are any hidden safe troubles, such as leakage of water, gas, electricity and potentials for fire accidents.

At that stage, will you still treat it as a tiny smart camera? To a degree, we can call it a smart home robot that is capable of deep learning. Such capability of deep learning has already been demonstrated in AplhaGo’s match with Lee Sedo.

When I sat that we are going to make commuting, people’s house, kids and the aged safer, we are not being sensational and talking nonsense. Instead, we will achieve these goal through R&D and acquisition and steadily develop our technologies and products. Qihoo 360’s goal is never to deliver takeaway food or sell tickets online. Instead, we are a tech company, and our goal is always to serve the general public via technologies and help people take better care of their beloved friends and families.

I believe now you can understand why we are not too much focused on O2O takeaway food and online movie ticket-booking services. Every people at Qihoo 360 should set American startups founders and engineers as their example and develop some really breath-taking and meaningful products one day. You should not be devoted to making more money, but rather to making something novel and solving problems for the entire human race.

Once the age of artificial intelligence starts, every one of you will be put into a new stage where technologies update exponentially and improve unstoppably. What you and me are left to do is nothing but seize the opportunity and embrace change.

By Zhou Hongyi

(Like our Facebook page and follow us now on Twitter @tmtpostenglish and on Medium @TMTpost and on Instagram @tmtpost_english.)

[The article is published and edited with authorization from the author @Song Changle, please note source and hyperlink when reproduce.]

Translated by Levin Feng (Senior Translator at PAGE TO PAGE), working for TMTpost.

本文系钛媒体原创,转载请注明出处、作者和本文链接
分享到:

第一时间获取TMT行业新鲜资讯和深度商业分析,请在微信公众账号中搜索「钛媒体」或者「taimeiti」,或用手机扫描左方二维码,即可获得钛媒体每日精华内容推送和最优搜索体验,并参与编辑活动。

tmtpost
tmtpost

Our official account in English/English Version of TMTpost.com

评论(0

Oh! no

您是否确认要删除该条评论吗?

分享到微信朋友圈