The age of speech giants is coming! Who will be China Siri?

Editor's words: Today, the Voice technology has become the standard of the Giants, Baidu, Tencent, Sogou, Cloud to know their voice or the traditional internet giants, or from the Chinese Academy of Science and technology giants. It is inevitable that giants will occupy voice highs. And from the beginning of last year's voice market by the industry's attention, to today's cloud to get tens of millions of dollars financing, voice market spring has come? Does the entrepreneur still have a chance? See what the author says. (By, Song Xuan) text/Luo (Sohu it exclusive release) about a year ago, China Mobile at the cost of 1.36 billion yuan won the hkust 15% of the stake, the latter stock price rose, from less than 30 yuan to the highest 61 yuan, to become a cattle stock, the market value of up to more than 24 billion yuan. And in just the past long vacation, "cloud know Sound" also high-profile announced to obtain the amount of tens of millions of dollars to about 100 million yuan a round of financing. Although its volume is not comparable with hkust, but this is silent for a year speech recognition market is a significant positive. And also involved in the field of Baidu, Tencent, Sogou is also speeding up the pace of technical iterations, as a strategic level, voice technology in the eyes of the Giants are particularly important. Similar to the foreign Giants ' occupation of the voice market, it is widely believed that the voice market startup window has been closed, subject to technical barriers. Only belong to the giant voice of the warring states curtain pull! Siri has burst into the domestic voice market. Since the launch of Apple Siri, the voice market has received high attention. People even exclaim that this more natural manipulation will replace the keyboard. For a time the followers came into the bureau. Apple rivals Google with Google Now strong kill, with search technology expertise and data accumulation, in the interactive effect of above. The much-watched Google Glass is started by the cool voice of OK Google. Domestic internet companies Baidu, Tencent, Shanda and Sogou are the introduction of voice-related products. Sogou last November launched a voice assistant, its voice recognition technology is the use of "cloud known sound." And the use of its own voice recognition technology Baidu, last Christmas during the launch of voice assistant, 1 months later than Sogou. Tencent in the voice market is much more conservative, micro-letter rich in voice walkie-talkie, can be naturally transplanted to voice assistants. But in addition to the introduction of the "Voice alert" public in 4.5, other speech recognition features are not enhanced. This is also consistent with the style of Tencent, in the model is validated, the market is educated to be able to exert force. It is noteworthy that the micro-letter built a technical team of more than 30 people to develop speech recognition technology. In addition to Baidu, Tencent and other giants, Shanda launched the use of its own technology, "Lark Voice assistant." China Mobile teamed up with Hkust to launch the "Connect assistant", the message fly itself and "language point" this voice assistant products. Entrepreneurial intelligence 360, Worm hole voice assistant, small I robot focus on semantic analysis and front-end functions. The popularity of speech technology has more restrictive conditions of speech even if Siri is still not the mainstream of interactive mode. There is no product that can be equated with "voice assistant" in China. China Voice market last yearAfter an "arms race", no one is sure whether the user really needs such a thing. But now, we all spare no effort, for fear of backwardness, be people occupy the legendary entrance. 1, "Voice entrance" may be just legends. I am not an afterthought, when Siri launched I think voice interaction has a natural flaw: only in a quiet scene to use, noise is difficult to identify, only in the privacy of the scene, otherwise voice commands will interfere with others. In reality this kind of place is not many, even if the living room in the home uses the voice function, may also affect the family member. Even if there is no interference, speech recognition technology has a dependency: wireless network. Upload a large number of data for cloud recognition, must have a good network. In places where there is no WiFi, using voice control is a nightmare. There are products to provide off-line identification technology, installation package will increase the number of times, the recognition effect will plummet. 2, the intelligent technology of voice products is still passable. The more difficult problem with the voice market is intelligent recognition. Speech technology is divided into speech recognition, semantic parsing and speech synthesis. Most voice search products can do is only to convert the voice to text, and then search through the text, is actually "speech recognition" This part of the technical application. The degree to which a voice assistant is an assistant needs to be able to understand the language, to comprehend natural languages, like the Jarvis System in Iron Man. Now the "Voice to text" step is not natural. and understanding the natural language, is still Google, Baidu and other technology giants in the difficult point: semantic analysis. 3, the user has not yet formed a habit finally there is a difficulty in user habits. Good products to touch users, need to cultivate, change and education. It takes time. At present, the use of scenes, wireless networks, semantic recognition and user habits make the voice still in the Pathfinder period. Hit a lot of resources, did not get a matching harvest, so there is a bubble. The time window of the entrepreneur is over! Today, speech semantics is a battleground, especially when wearable devices emerge. The voice market is bound to become the giants of the game, technology and data threshold high, and voice business time window may have passed. 1, the giant transformation speed is accelerating. The role of "cloud" and "hkust" in these contenders is to provide data and technology for use by top service providers, just as in the Map field. Coincidentally, the hkust is also seeking to fly from "B" to "C" the diversification of the road of transformation. In addition to voice assistants, Hkust also launched a flying voice input method, for preschool education, voice robot hardware and other products. Diversification is now what speech giants are doing and what speech recognition platforms have to do. They offer free identification technology for the enterprise market, but custom fees alone won't work. The internet giant has always been keen on free, open platform routes, netting developers to gain access to traffic, data and personal users. At the end of August Baidu navigation completely free, and the direct confrontation with the gold is a living example. Baidu, Tencent and other companies are investing huge sums of money and resources to strengthen voiceTechnology construction. If they will be free speech recognition technology will inevitably create a larger voice biosphere. 2. The advantage of resource technology has become an obstacle to entrepreneurship. In fact, the basis of voice business is to build it in strong technology driven to achieve, and currently only the Giants have the relevant advantages. At the same time, voice technology plus semantic parsing technology, with the help of knowledge map, in-depth learning, to achieve a dialogue search, in the mobile internet era can explode a great energy, and more easily landing and commercialization. It is a big problem to simply provide speech recognition technology and make a technology platform. and Baidu and other internet giants in vertical integration compared to the professional field of entrepreneurs, more advantages. is the voice market spring coming? Perhaps these problems will persist for a long time, but it is undeniable that voice as the frontier technology of mobile Internet is still worth the industry's expectation. In terms of usage scenarios, Glass's "OK Google" is a start. Music video TV, Hammer OS, Ishin, Inwatch, millet 3 and other products have launched Voice interactive function, are used cloud to know the sound or fly this two company technology. The rise of wearable equipment, the wave of hardware entrepreneurship will bring more of the use of voice interaction in the soil. For environmental noise interference, Baidu is responsible for multimedia search technology, Dr. Yukei a few months ago to the author explained the anti-noise technology, speech recognition technology has evolved to distinguish between human voice and environmental noise, or even according to the voice of a person to identify a specific sound. This technique can also be applied to mobile payments. Alipay has launched "sonic payments" using sound fingerprints. This also shows that the use of speech technology will only be more and more explored. For example, corporate customer service. Today, there is news that the hkust and Anhui Mobile signed a large list of nearly million. Relatively mobile, tens of millions of scale is not a big single. But their collaborative content or will trigger new business call center upgrades: Later China Mobile in the customer service 10086 platform will use the intelligent voice technology, users can speak directly to demand. The voice of customer service will also use voice synthesis technology, which is the advantage of the hkust, its voice synthesis can even support the mainstream dialect. Affected by this news, today's HKUST news fly stock trading. The environmental improvement of wireless networks is also good news. The 4G license plate is already an arrow on the string and has to be sent. The author recently got China Mobile 4G (LTE) online card test shows that 4G network in the single user use bandwidth has been as high as 44M, download speed of 4mb/s. This speed will be affected by the number of users, but can be predicted 4G compared to the 3G era is not the same. In addition to 4G, the enthusiasm of operators, governments and businesses for WiFi has also brought a broader range of wireless hotspots. Finally, who will become China's nuance is still inconclusive. But this market has brought many practitioners unlimited imagination space, can be foreseen, the future voice market will be in the giant's scramble to become the industry focus, and downstream voice products will gradually enrich, a mobile voice ecological ecology or will soon form ...