A journey from P20: How AI can change the future of mobile phone life

Spring comes, with all the initiation of all things, there are all kinds of mobile phone new ...

Compared to last year's full screen trend, this year's mobile phone market can be said to be AI world. A variety of different levels of AI photography. The advent of the  P20 series has also lifted the topic of mobile AI to new heat. For example, based on the npu of the Kirin 970 powered by the AI special processing capabilities, P2o for the first time to achieve the AI photographer system, in the photography to give users an instant master experience, and AIS intelligent anti-shake system using the NPU provides image recognition capabilities, imaging quality significantly improved It solves the problem of shaking our hands when we take pictures and shoot videos. However, it may be more consumer's impression that AI is a very powerful, similar terminator to the existence. Even today, we can see that AI can do all kinds of strange things. Come to the mobile phone, can ai only be used as a photographic aid?

Of course not. In many of the mobile communications and AI experts we interviewed, the ultimate goal of mobile AI is clear: to make the phone feel and think like a human. This road will not be done overnight, but the open gate will certainly not close. The AI photography We see today based on the AI special processing chip may be considered as a beginner version of the future of mobile AI: the mobile phone has its own AI, has the ability to recognize and perceive, and is able to make basic judgments and inferences about its own external objects.

So let's blow our brains out and imagine how AI will change our mobile life over the next few years based on this model.

Apps become less important as natural interactions are pervasive

Mobile app This thing is definitely a reverse-day invention that changes the lives of all mankind, but it seems that from the date of its birth, people are thinking about how to cancel it. After all the dazzling app behind, Hidden is the life is severely fragmented. The switching interaction between apps also seems to be the "time killer" around us.

Can this model be changed? Or is there some way that we can speed up the use of mobile phones, rather than spending time between endless switching apps? From the innovation of  P20, this possibility is becoming a reality. Taking photographic technology as an example, in the past, because the mobile phone terminal lacks specialized processing picture module, wants to obtain a picture which satisfies oneself, must depend on various picture processing app. This approach is a step trouble, and the second is that the app can only be processed based on the accuracy of the original image, resulting in a drop in the quality of the image. The development of P20 's intelligent graph function can make the judgment of portrait, landscape, Night scene and so on when taking pictures, so that the parameters are processed directly. This is to get rid of the image processing app dependence, let the user see is the gain, one second to become master.

For now, this is probably just the beginning. With the voice recognition, semantic understanding, the further maturation and deployment of NLP technology, and the continuous enhancement of physical recognition through machine vision, the AI system in mobile phones strengthens the learning ability of users ' habits and service needs. The combination and optimization of these AI technologies with mobile phone products makes it possible to control mobile phones in a natural interactive form. So we can from the "app waiting to be point" into "mobile phone to provide services," the new model.

Imagine if your phone can recognize your voice print in your pocket, whisper that I'm asking for a car, and come to the car right away, say I'm hungry, and immediately follow your taste and physical condition to provide food and beverage options. Or in another case, when you take the phone out to the product, the cell phone can buy it home, and on the landscape, get a big picture immediately.

Some experts say that in the future mobile phone life, in addition to social and information interface, mobile phone only left a hearing and a visual entrance. Mobile phones can turn complex services into simple interactions by approaching human auditory, visual, and understanding capabilities.

AI-based mobile phone "de-application", can be seen as a mobile phone like human, to save human mobile phone time, and even more than the current mobile phone can not achieve the application function.

Of course, all this is premised on the abstraction of services and the sharing of data between apps. By then, businesses will not compete for the use of information to occupy the user's time, but by direct advice and communication, affect user spending habits.

physical world recognition brings fantasy imagination

In the  P20 series, we have seen the magical experience of using mobile phones to identify the physical world, such as identifying cats, dogs, landscapes, and portraits for photography. When it comes to identifying landscapes, AI photographers use AI algorithms to process, analyze, and understand the image data captured by cameras and sensors so that they can clearly capture the landscape. Then through the more complex algorithm learning mechanism, the image features are captured and processed to achieve optimal adaptation to the scene.

In addition, P20 also has the ability to recognize characters, identify and restore user jitter, and even identify and learn the user's usage habits.

It is not difficult to see, AI mobile phone most of the functions, even can say more than 90% of the function, are recognized. This is based on the theory of machine learning, the basic performance of AI development: through learning and extracting physical world characteristics, to continuously enhance the recognition accuracy, to achieve the traditional computer can not complete the task.

Interestingly, Ai's ability to recognize, even if focused on a mobile camera brings recognition, is far more than just the recognition objects we see today to help take pictures, and recognize text to achieve translation. In fact, the recognition of all things is only a matter of time in the eternal superimposed machine learning tower.

What will we recognize in the future with mobile phones? That's a lot of brain holes.

For example, when traveling, we may not know a building, a monument, a symbol, then we can take out a phone shot, and then the knowledge and information behind the knowledge map into our front. Today we went out to the outing has been accustomed to the flower pattern. But this is just the beginning, in the future we may not need a tour guide, no translation, no driver, a mobile phone easy to fix the world.

Another example is in the diet. We are only making food today just to make a circle of friends. And then, we seem to be able to take pictures to identify food practices, nutrients, collocation rationality, and so on. Even into a group of "dark food", can also be based on the approximate degree of food, food products, such as the ability to restore a glance through the AI, to remind users whether an unhealthy diet. Of course, all dietary advice should be based on the understanding of the individual, so that mobile phone learning user health data, record each meal nutrition curve, seems likely to become our future life style.

There is a lot more to this from AI recognition, such as we can tell by AI how many things weigh, how big a room is. A lot of work in the measurement and weighing work can be laid off, only with a mobile phone.

such as AR need to rely on high-precision environmental recognition, which is the future of AI is one of the main direction. When the ability to recognize the environment in mobile phones is further enhanced, the AR and Mr ERAs may begin formally.

The scene of the ubiquitous recognition, multi-layer recognition, today seems to have almost infinite imagination space. Mobile AI technology and countless clever brains will be the "sci-fi movie" to become the real-life.

experience, wisdom, and knowledge can all be transformed into models and even commodities

In  P20, we saw the AI Master System. The system integrates complex algorithms and models to refine the knowledge and experience of a real-world photographer into a solution that allows small white users to shoot top blockbusters in minutes. For example, when the camera detects complex scenes, such as flowers and people in the same frame, the AI master will give a logical priority to the photo solution, giving priority to the intelligent shooting scheme for more important subjects. This is similar to the human photographer's experience and judgment as an AI model, so that small white users can also be like a master to shoot a satisfactory "high-sense" blockbuster.

We may as well open a brain hole, is only the master photographer's wisdom and experience can be compressed into AI? Is it possible for other people?

In fact, our invaluable "experience" in life, if broken down into mathematical language, is probably the fuzzy control, correspondence and priority of various parameters. But the human brain does not have these mathematical concepts, but can turn them into a paradoxical feeling.

In the AI world, the machine learning algorithm itself is the abstraction and reuse of the function of "experience". When this pattern comes to mobile phones more and more, all kinds of experience, wisdom and knowledge are abstracted and become AI models, and become a part of our life, it seems to have become possible.

For example, today's girls will look at a lot of beautiful makeup, learn how they choose beauty products. This often requires a lot of study time, not necessarily the right choice. If we ignore the happiness and achievement factors in this process, simply abstract the experience and judgment of the beauty-making master into the AI system, so that consumers can choose the AI that suits them, and use it to choose their own products, then it seems that everything becomes more scientific, time-saving and more in line with individual requirements.

When we are photographing, we can use the experience of the photographer (and even switch between the different photographers and artists); When you cook, you can take a foodie's style, and when you read a master's taste, the phone may become a new human enhancer, and we will save a lot of learning costs, But still enjoy very high efficiency and high quality experience and aesthetics.

All this, from AI and AI photography, if Mate10 and Kylin 970 is the initial encounter AI, then the P20 series we can say is to witness the growth of AI, witnessed the AI as a new way of interaction, value provider, to the user to reach the possibility of service.

Although today is a long way from the future AI life we envision, the floodgates will not stop once they are opened. The rest, however, is the time to enjoy a technological revolution.