Do you have a lot of artificial intelligence TVs on the market? Are they really smart?

Have you ever thought about the question: Why does the artificial intelligence plus industry seem to have infinite brain holes, such as AI medical care, AI finance, etc. Once you enter the artificial intelligence plus hardware, what will be wrong? For example, an AI refrigerator, an AI washing machine, an AI light bulb... There is a strange feeling of guilt in sketching that thing in your head.

In principle, this is because most of the hardware products have been strictly defined when they are designed. Light bulbs are light and the washing machine is responsible for washing clothes. Putting these things on the CPU may bring about a good interactive experience.

However, artificial intelligence relies on data collection and deep learning, and common home appliances do not have access points for these things.
In other words, it is difficult for household appliances in the general definition to provide AI with the basis of the “cognitive-judgment-decision-making” three-step process, and to link the artificial intelligence and hardware together in the spirit of the Internet that “it is difficult to create difficulties without difficulty”. It may bring something very embarrassing - for example, artificial intelligence television.

In the spring of 2017, quietly passed. It is AlphaGo's magic that most people may remember. But silently, a large number of artificial intelligence televisions have been accumulated in the market.

The question is, do these TVs really have artificial intelligence?

In spring, a large wave of artificial intelligence TVs has been awake

According to industry practice, the spring is a concentrated time period for major TV manufacturers to release new products. In the spring of 2017, users and media concerned about the television industry can really recall the “fear of being dominated by artificial intelligence”.

In fact, since the second half of last year, Changhong has released the Q5N, which focuses on artificial intelligence, and is equipped with a television-centric artificial intelligence platform called AI Center. Immediately afterwards, almost none of the TV players fell. For example, in October, Xiaomi released the 3S series of artificial intelligence televisions.

When time entered the spring of 2017, the situation could not be controlled.


At the end of March, TCL introduced three TVs equipped with an artificial intelligence system “Little T”; Lenovo’s spring conference launched Lenovo’s 17TV 65i3, which focuses on intelligent voice interaction; Xiaomi’s 4A series launched at the spring conference also continued to focus on artificial intelligence cards. , And created a new voice system PatchWall.

In April, Microwhal launched the Micro Whale Smart Voice TV 2.0 product, and announced that the full line of micro whale products will also enter the 2.0 era. In May, Storm launched the artificial intelligence TV X5 ECHO. Just a few days ago, it gathered the research and development achievements of Haier and Ali. The fourth generation of Haier Ali TV also uses artificial intelligence as the main publicity point; in addition to the LeTV Super Video, which has been playing a major role in voice interaction, all domestic TV players are basically “artificial intelligence”.

This scene is very explosive: There was no AI TV on the market half a year ago, and there were nearly ten brands of AI TVs on the market half a year later. Artificial intelligence and the madness of the 3C industry can be seen from this.

At the same time, the problem has also come. So many televisions, in the end how "artificial intelligence" it?

Three-legged paradox: All AI TVs can't escape the cage

We know that although the definition of artificial intelligence is widespread, the basic boundaries are very clear. That is, the machine computing system operates in a way that simulates human intelligence. This involves the basic process of "cognitive-judgment-decision-making" mentioned above. This kind of operation can effectively reduce the amount of computation and achieve better interaction with humans. To say farther away is to give the machine the unique capabilities of organic intelligence, such as intuition, reasoning, and judgment.

But today's AI TVs around us are obviously not as high. Looking at all the AI ​​TVs that have been launched, there are only one and two main modes of reflection in the independent field of AI: the main one is voice control, and the second is smart recommendation content and home system control.


Analyzing the above-mentioned AI TV products, we can find that all products are mainly voice control systems, emphasizing the recognition and diversity of voice control (such as simple voice question and answer, voice activated calendar and other non-video applications). Smart recommendation is also the core of AI TV. For example, Storm's AI TV products emphasize thousands of people and more and more understand you. Millet's AI TV has smart wall function. In addition, Haier Ali and other artificial intelligence televisions also use TV to control home appliance systems.

In addition to these three, there are also some AI TV announcements that can be extended to other AI applications. For example, Xiaomi demonstrated face recognition at the press conference. Micro whale's AI TV is said to have emotion recognition and other functions in the future. However, these functions are basically still in the experimental period and data. I am afraid there is still a long way to go to real applications.

However, as far as the three main functions of AI TV are concerned, the proportion of technology applications in the artificial intelligence core fields involving machine learning and intelligent identification is not large.

Take the AI ​​TV's voice control function on the market today. At present, the so-called AI TV voice interaction is generally divided into two kinds, one is through the voice directly activate the command, such as "shut down" "put a certain movie." The other is search-based speech interactions, such as "I want to watch comedies" and "I want to see someone's movies". These functions are based on human-computer interaction in classical computing. The difference is only the size of the database and the sensitivity of the reaction.

The components involved in semantic understanding do need to use artificial intelligence algorithms to generate derivation results. However, current AI TVs are still weak in terms of true semantic understanding. Often only part of the problem can be solved by keywords. Complex issues and expressions after word replacement are unacceptable.

The number of AI components applied is only one aspect. The more important issue is that these voice interactions do not give consumers any substantial improvement. For example, in long-distance speech recognition, family multi-person speech recognition, and crucial dialect recognition, the current solution offered by AI TV in the market is still not ideal.


The smart content recommendation actually faces such a problem. Under the current TV hardware system, smart recommendations can only recommend content based on your browsing history and viewing duration. This actually uses only one key word retention and data analysis technology, which is far from AI. More importantly, the current AI TV smart recommendation does not solve the problem that different family members want to watch different programs. And, watching TV is different from the purpose of watching a mobile phone, and content recommendation affects the user's viewing experience.

However, it is very far from being able to control households through TV. It can even be said that it is nothing more than putting the home remote control on the television screen. While controlling the refrigerator washing machine water heater through the TV screen .... Oh .... really a lot of people need it?

Road to see a roar: why suddenly AI TV?

AI TVs can spring up like mushrooms. I am afraid it is difficult to explain them with many vendors. There is only one conclusion that can be accepted: The new concept product launched by the friendly trader has a low threshold, and can rely on the existing technical strength to imitate or even surpass it immediately.

But the AI ​​laymen in many television circles suddenly entered the AI ​​in a neat and tidy manner, and even started a homogenization battle in the short-term, which is by no means without reason. Some of the problems that have accumulated from the TV industry may explain why a large number of AI TVs were airborne in a matter of months.


First of all, it is the interaction of television that has persistent problems. In the era of Internet TV, the content and playing format that users can choose have experienced a sudden surge. The traditional TV station model was quickly abandoned. However, with the rapid increase in content selectivity, the problem of human-computer interaction has not been resolved. Smartphones can use touchscreens to complete human-computer interaction iterations. Can't the TV always poke on the big screen?

As a result, various technologies have come one after another, and some use mobile phones as interactive portals. However, the experience is poor and compatibility is very problematic. Some have introduced new Internet versions of TV remote controls, but they are too expensive and do not meet user experience.

After repeated sawing, the consensus in the industry is that voice interaction should be the best solution. After all, remote control has been popular for decades, and other methods are difficult to replace in the short term. The voice input is human instinct, perhaps to be able to better complete the manipulation of massive content - just this product innovation and AI's earliest mature speech recognition technology, the conclusion is of course together.

Second, it is difficult to incubate new market excitement within the traditional track, and the television industry needs new concept stimuli to complete self-salvation.

In the first quarter of 2017, the total sales volume of color TVs was 11.38 million units, a year-on-year decrease of 13.5%, and retail sales totaled 36.6 billion yuan, down 7.3% year-on-year. The overall weak market was driven by OLEDs, quantum dots, ULEDs, and laser displays. The self-iteration of the picture quality and ultra-thinness was too many times, and even caused the consumer's aesthetic fatigue. After all, especially good image quality and large screens are not necessarily needed by every family, and even most consumers do not seem to have any difference.

It can be said that television entered the era of homogenous competition earlier than mobile phones, and the concept of artificial intelligence may create new trigger points for the market. Anyway, idle and idle, R&D costs are not high, and TV manufacturers are not happy?

Once again, the future imagination brought on by artificial intelligence has also led to a congregational competition. Today, AI is a technological singularity that can be expanded almost infinitely. What kind of revolution will be brought to the television and the entire home system afterwards is still not easy to judge, so early entry into the market has become the choice of many businesses.

The final conclusion is that we have seen a large number of AI TVs with similar capabilities, similar appearances, and poor PPT. The entire vertical industry presents a prosperous situation momentarily: The road is not flat, and you have everything I have...

Do we really need artificial intelligence TV? I think it should be ... need ....?

The main judgment is based on the fact that TV, as a multi-terminal terminal for collecting voice, video and video, must enjoy the mature technology of AI. On the one hand, as one of the terminals that interact with humans for the longest time, AI can provide a great imagination for human services.

There must be some stories about the addition of the two, but the biggest hidden danger in current gameplay is that it is likely to give the stigma to the concept of artificial intelligence television. By the time the artificial intelligence technology matures on the TV, both the market and capital will lose their enthusiasm.

The story of the development of the seedlings and the Chinese Internet circle has not only been heard once or twice.

What AI TVs do you need to go through?

Perhaps in the future, we can only continue to follow the footsteps of manufacturers' PPTs and public relations papers, and call those more mature AI TVs "real AI TVs" or "AI TVs unmatched" and more exaggerated names. But real AI TVs may need to complete substantial breakthroughs in the following areas:


First, multi-level complete human-computer interaction: voice control Even if a very high level of in-depth learning is achieved, judging the user demand for the AI ​​TV is relatively single. Because voice data is mostly a direct command from a user to a machine, it is difficult for AI to achieve a three-dimensional understanding based on it. When the overall user’s body, face, emotion, and other data sources reach an overall judgment, it is more likely to achieve a complete leap in the overall user experience. But this requires the first evolution of television's own sensing device.

Second, AI understanding of the content: When we add various functions to the TV, we tend to ignore the essence of video is the video content. However, if AI has its own understanding of TV content, it can make in-depth judgments based on the understanding of the content. The release of information to the user in the reverse direction, such as recommendation, recording, and viewing methods, should make the use of AI TV truly iterative.

For example, when you are watching a movie, AI will tell you more about the effects of brightness, contrast, and sound effects, and even if you decide whether or not you should open the window for ventilation, it is obviously different from TVs that can only recommend programs based on keywords.

Third, function integration and AI control: TV's intelligence is based on the complexity of its operation. Various functions, frequency bands, and external devices make TV a responsible comprehensive terminal. But this is obviously inconsistent with the nature of television. If the AI ​​can understand the connection between various devices and functions, intelligently manipulate the TV, obviously can be somewhat different.

These changes require a process, but most of them are not difficult to reach. But at least a lot of R&D and product investment is needed, which may also be the scarcity of the impetuous AI TV market.

Which company is most likely to win this battle?

This is really not known.

But the most likely outcome of this blitzkrieg is: rushing horses and losing all the games.

Smart TV/box information can focus on smart TV information network sofa butler (http://), China's influential TV box and smart TV website, providing information, communication, etc. on TV boxes, smart TVs, smart TV software, etc. Answering questions.

Creative led display screen

LED DJ Display

LED DJ Booth is used different product models and structures as different conditions, which can be customized to the scene modeling. It can be spliced out the unique shape with different sizes of triangles, rectangles, hexagon. Unique DJ Table is not limited to kinds of environment. The application of the relevance and the shape of DJ LED Video Wall can enhance scene recognition.

LED Bar Display will not only present a delicate and realistic color performance through the synchronization adjustment, but also is compatible with different sources of input such as TVV, HDV, DVI, VGA, SDI. We adopt SMD technology to achieve super-wide viewing angle and better surface smoothness. Super light cabinet is to realize easy installation and maintenance that can be customized for shapes, sizes, pixels, resolution.



LED DJ Display, LED DJ Booth, DJ LED Video Wall

Shenzhen Priva Tech Co., Ltd. , https://www.privaled.com