Can IBM the first player to beat a human player be able to get a start in AI

Source：Updated：2017-09-11 13:05:11

With net profit falling, share prices slipping and Mr Buffett down, IBM's transformation has been tough. It is fairly common to say that IBM is missing out on the AI era after missing the mobile era. 
 
Indeed, the hardware is overshadowed by nvidia's GPU and Google's TPU, and the algorithm is much larger than DeepMind's huge paper, and its application is much broader than Microsoft's and baidu's. IBM was somewhat awkward in an age when everyone was holding up AI. 
 
To say that IBM has done nothing on artificial intelligence is wrong. Maybe have you ever been baidu, sogou robot brain impressive performance in variety shows, so you should know that, as early as in 2011, the IBM supercomputer Watson in the quiz show jeopardy beat two human players, with history. 
 
From then on, or even earlier, Watson became a force in IBM's battle against AI. 
 
IBM Watson core capabilities: unstructured data 
 
Watson is based on the name of Thomas j. Watson, the founder of IBM, who is accurate to say that Watson is a "platform" or "system" capable of understanding, reasoning, learning and so on. Since the beginning of the last century, IBM has been exploring the field of artificial intelligence, and Watson, who can communicate and refine information, is only one of the results. 
 
Not to mention that even the game of chess and chess has been played by IBM before anyone else gets on the quiz. In 1997, deep blue, an IBM computer program, overcame kasparov, a chess player, in the field of chess. 
Can IBM, the first player to beat a human player, be able to get a start in AI? 
From dark blue to Watson, it is IBM's most critical choice in AI direction. 
 
Some people say that deep blue is Watson's predecessor, and that's not true. Deep blue is a game of chess, which is a number game, and deep blue also has 480 chips specially designed for chess - most likely to be violent and poor. Watson's application is in the man-machine conversation, which deals with language, sounds and even pictures. In contrast to the deep blue hardware, Watson has been unshackled and evolved into a technology platform. 
 
The key point is that Watson is focused on understanding non-structural data. 
 
To understand what is unstructured data, you need to know structured data first. Structured data exists in ERP system and data report. It is the data that is annotated and exists in the database. So unstructured data, nature is what you can only use. TXT,. MP3,. Avi and other formats store data. 
 
So Watson's understanding, analysis, refinement, and reasoning can be built on natural language and images. 
 
Cognitive computation and artificial intelligence are differences in language skills and technical differences 
 
So Watson is more of a giant, bovine data-processing platform, but in an era when everything is based on data, which artificial intelligence product isn't? 
 
In a distant 2011, IBM seemed not ready to put Watson on the artificial intelligence vents. Instead, IBM came up with a very fresh concept for Watson: Cognitive Computing -- Cognitive Computing. 
 
Cognitive computing explanation is this: "cognitive computing represents a new computing mode, it contains the information analysis, natural language processing and machine learning in the field of a large number of technical innovation, can help decision makers from the large amounts of unstructured data reveal remarkable insights. Cognitive systems can interact with humans in a more natural way to humans; The cognitive system specializes in the acquisition of massive amounts of different types of data, based on information; Learn from your own and your data, your interactions with others. 
Can IBM, the first player to beat a human player, be able to get a start in AI? 
So it's really hard to tell the difference between cognitive computing and artificial intelligence, especially weak ai, and the definition is very similar. It's all based on data, using NLP, machine learning, and so on to give machines the ability to learn and evolve, so that they can better serve humans. 
 
In Google's search for Cognitive Computing, most of the results are related to IBM, and it now appears that everything about Cognitive Computing seems to be a failure of IBM's word-building program. 
 
Was it because of the inadvertent choice of speech, that IBM lagged behind in the AI era? 
 
Watson's report card was not as bad as it was supposed to be 
 
Regardless of the technical concept, IBM's cognitive computing is somewhat different from the artificial intelligence that many companies now dabble in, and these differences exist in the market jurisdiction. 
 
Companies such as amazon, Microsoft and others will fall to a specific level of civilian production -- Echo, ice, tesla... While it is not clear where the target markets of these companies are, they are all in the c-terminal market. 
 
The same reason that everyone else followed the artificial intelligence box, you don't do it, it's like you're not in this field. 
 
Perhaps from the genes of consulting management, IBM has targeted the enterprise market at the beginning. On IBM China BBS in 2014, IBM offered to "help transform the industry with data" as one of the main areas of transformation. The logic is simple: cognitive computing is based on data. IBM has a large number of enterprise data, and naturally it has launched a service that applies to enterprises. 
 
And IBM came up with a handful of hard bones: healthcare, finance and marketing, but Watson's performance was not as bad as some analysts had said. 
Can IBM, the first player to beat a human player, be able to get a start in AI? 
In the medical field, Watson was the most vocal and vocal. 
 
Training Watson for cancer treatment, and the Anderson cancer center, and so on, has attracted a lot of media attention. In the medical field, Watson has the greatest advantage of finding and refining the data, and Watson is said to be able to read 800 million pages of information per second. It's just beautiful to be able to quickly process data and assist doctors. In fact, five years after collaborating with the Anderson cancer center, the two sides failed to produce a practical tool, and eventually the Anderson center paid reparations and ended the cooperation with IBM. 
Can IBM, the first player to beat a human player, be able to get a start in AI? 
(IBM concierge robot) 
 
In contrast to health care, Watson's path to finance is much smoother. 
 
The introduction of the bank branch service robot, the release of financial cloud, auxiliary intelligent investment, and so on, although mostly still some small work, but the ride of Fintech this trip must not go wrong. And IBM was a big player when it came to acquiring financial firms, with Promontory and Algorithmics in the past, and probably more in the future. AI+ finance has become a big trend, and with IBM's close relationship with Banks, it may be Watson's home. 
 
In terms of marketing, Watson has a broader scope. 
Can IBM, the first player to beat a human player, be able to get a start in AI? 
AI+ marketing is a topic that is more or less commonplace. Watson is said to have learned nine lessons each day from IBM's thinking on marketing programs, with the initial goal of "knowing how to tailor creative content to every customer." The goal is big, but what Watson is doing now is the same old story -- learning user history and consumption data and making intelligent recommendations. Last year, Watson teamed up with The Northface to recommend a suitable coat for consumers through The user's browsing history, local weather and other data, VentureBeat statistics, which hit 90% of The time. Compared with medical and finance, AI marketing has less restrictive and high fault tolerance, which can give Watson more room for development. 
 
Born in innovation, dead in sin? 
 
IBM today, in fact, comes from the original sin of the B - end service. With a large volume and stable revenue, it is difficult to transform even the AI market. There is a natural limit to the data that has been served in the B market, and the purchase of data in the c-terminal market/research and development of B terminal AI technology has two choices in the balance. IBM's choice seems like an easy path: using the original data to develop the product and then allowing the long-term collaboration of the enterprise users to pay for it can be seen in practice as less easy to buy. 
 
And the long term investment in the B market, the unflattering "cognitive computing" talk, the consequence of IBM is to miss out on the golden age of mass mind. But IBM's technology is still there, and even if Watson's current performance is mediocre, it won't be hard to see the future with Google, Microsoft, and amazon. 
Can IBM, the first player to beat a human player, be able to get a start in AI? 
 
Needed to be aware of, but the domestic many B side service enterprises, transformation without IBM's ability also got IBM's disease, think the enterprise market will always be welcome, even the most basic technological innovation ability have been lost had been sinecure. In an era of rapid artificial intelligence technology, any step in the wrong step can be a dead end.