With the mature application of big data, cloud computing, artificial intelligence and other technologies, the children's intelligent industry has become a hot industry in the current wave of intelligence, and the capital of all parties has entered into it, and the industry has entered the fast lane. According to relevant data, the market size of China's children's consumer industry is 2 trillion yuan, and the scale of children's smart products represented by children's smart watches and children's intelligent robots reaches 100 billion. The future development of the industry is occupied by the industry. I am optimistic.
In this context, the "2019 China Children's Intelligence Industry Annual Conference" hosted by the China Intelligent Terminal Industry Conference Organizing Committee and Shenzhen Children's Intelligent Products Association and Shenzhen Children's Intelligent Products Association was successfully held in Shenzhen. At the meeting, not only the “Top Ten Brands of China's Children's Intelligence Industry” but also the “Top Ten Potential Brands of China's Children's Intelligence Industry” were selected. During the meeting, the editor of OFweek Weike was fortunate to interview Dr. Kang Heng, vice president of Beijing Yunzhisheng Information Technology Co., Ltd., and there is another identity behind Kangheng, the vice president of Shenzhen Children's Intelligent Products Association. For this conference, Vice President Kang Heng said to the editor of OFweek Weike.com: "This China Children's Intelligence Industry Annual Conference will comprehensively summarize the 2018 market issues and learn lessons so that everyone can go better in 2019. At the same time, Kang Heng said from the perspective of Yunzhisheng: "As a technology supplier, we will share the ideas and judgments of the market with our partners so that our partners can be timely, effective and reasonably planned. My own products."
Yun Zhisheng Vice President Kang Heng
The primary stage of the children's robot market
When any industry is just emerging, it will not be a sound Industry, and children's robots are also true! Throughout the children's robot market, its rapid development is beneficial to the development and application of AI. However, for AI, although it has been recognized by people in the industry, AI is still in the development stage at present, and it cannot be applied to all walks of life on a large scale. It can only be helped in specific industries or scenarios. Users are more convenient and faster to improve work efficiency. The same is true for the
children's robotics market, which was established only a few years ago. For the global market, there are no relevant normative documents or experiences to borrow. As Kang Heng said: "Why the market is not perfect or immature, because our products are not yet subdivided.To provide different products for children of different ages for the user age group. The needs of different ages are not the same. It is not realistic to use a unified product to deal with all children. "
Legally speaking, people under the age of 18 are children. But for children's robotics, this range is wider. From their perspective, children are divided into five stages: 0 - 1 year old, 1-3 years old, 3-6 years old, 6-12 years old and 12-18 years old, children of different stages have different products and functional forms.
0-1 years old: due to At this stage, children's personal consciousness has not been established, and children's robots cannot be used effectively. Most children's robots mainly play nursery rhymes and watch pictures.
1-3 years old: Children at this stage belong to the exploration period, generally have certain Language ability, its children's robot can play stories, music or view pictures, and even play simpler poems for preschool education;
3-6 years old: children at this stage are in a fast-growing stage, not just Listen to stories, see pictures, and add knowledge and content such as encyclopedia questions and answers. If necessary, you can provide some hobbies and training for children.
6-12 years old: Children at this stage need more For complexity, it has to Beginning with a variety of educational issues such as language, mathematics, English, and nature;
12-18 years old: children's needs are more advanced at this stage, and existing children's educational robots may not be able to meet their needs and need to be re-established. Define or more specialized robots.
Kang Heng said to the editor of OFweek Vico: "What kind of product is a child robot? What kind of service should it provide? What kind of interaction should it have? There is no consensus on these, not only the company does not understand, many users do not know where their needs are. "
(1) Enterprise: Touching the stones across the river
From a business perspective, the market does not allow companies to consider perfection before entering the children's robot market, especially in this newly formed market. In order to better seize the opportunities, a large number of enterprises directly enter the market in an environment where conditions are not mature, and there is no relevant research and discussion on children. This also leads to the enterprise after several years of After development, there is no experience to learn from, nowWhat the children's robotics company is doing is to develop a large number of children's robots with different shapes and functions.
(2) User: What do I need?
is also true for users. There was no child robot industry a few years ago, users are not sure what their needs are, there is no way to tell which functions are needed by children.
In the face of the company's various forms of children's robots, users can't effectively distinguish how big the child is suitable for children, what can be done for children, but the user's own needs can not be clearly fed back to the enterprise. It will only lead to companies offering more ambiguous products.
Although children's robots are still in the primary market, with the development of the market, children's robot product forms and user perceptions will continue to move forward. Looking back at the children's robot market, if the children's robot has only one microphone and a simple dialogue a few years ago, now the children's robot not only has multi-modal interactive functions, but also adds various types of sensors, so that the robot can carry out some Simple motion function. The user can then perform partial programming, allowing the robot to do relatively complex movements such as dancing, kicking, playing the guitar, and the like.
Speech Recognition: Creating a more accurate voice interaction experience
On the children's robot, voice interaction becomes the most critical part: if the robot can't "understand" the user, then How will the next move be opened? But for now, most companies are not concerned about this issue, which is one of the reasons why the market has not yet unified product awareness.
In the experience of children's robots, children's robots often have unrecognizable user language, which leads to various problems such as unrecognizable, unanswerable, and incorrect answers. These problems have a certain probability of causing children. Cognitive error. Kang Heng, as vice president of Yunzhisheng, was invited to join the Children's Robotics Association and was one of the reasons for the vice president. Yunzhisheng's technical strength and research and development capabilities in the field of artificial intelligence.
Kang Heng said to the editor of OFweek Vico: "Cloud knows that it can be responsible to all users, Yunzhisheng can achieve 95% accuracy in speech recognition, and this is the national first line. The standard line of the brand. Secondly, the answer content of the device, Yunzhisheng has already targeted children.Research and development in this area has achieved very high accuracy. "
However, Chinese expression does not rely on a single word, but also involves the voice, personality and other factors of the speaker. In people's daily conversations, the dialogue semantics in different moods and environments are not the same. And whether the children's robot can accurately identify these conversations? Kang Heng pointed out: "For the moment, this answer is relatively pessimistic, mainly because the progress of technology has not yet reached this stage. "
The existing voice interaction mode is mainly intercom in a single environment, and can not effectively identify various dialogue modes and environments caused by emotions. Although many research institutions or enterprises are constantly researching and developing, the accuracy is accurate. The problem has not been effectively improved. However, the prior art has been able to achieve the level of recognizing the voiceprint, and to identify the identity of the speaker through the characteristics of different people's voices. In addition, the Yunzhisheng multimode demonstrated in the speech of the Kangheng Annual Conference State technology, that is, multi-modal technology capabilities such as machine vision and emotion recognition in addition to voice interaction will become the key to the breakthrough of children's intelligent products in the next stage. In November 2018, Yunzhisheng's image AI capability has reached the international first-class level. And has entered the development process of multi-modal chips. By integrating machine vision and speech recognition, it provides users with more levels of interaction mode, which makes artificial intelligence products more in line with user expectations and ultimately provides users with better experience.