Last night, Lei Jun held a live broadcast of the "Millet 5 Black Science and Technology Experiment" on the Xiaomi live broadcasting platform. In addition to introducing mobile phone performance, Lei Jun focused on demonstrating to fans the built-in “sweeping†function of the MIUI 8 system. According to Lei Jun's demonstration, the user only uses the mobile phone to scan the content of the subject, and the mobile phone can identify the content of the question and give accurate answer steps to help the students quickly learn the knowledge points that they do not understand.
(Scanning function demo screenshot)
However, the smart answer is just one of the “sweeping†smart applications. Through product screenshots, we can also find other applications such as “sweep code, photo purchase, translation, clapping, scanning business cards, and scanning documentsâ€. Industry insiders reacted to Lei Fengwang (searching for "Lei Feng Net" public number attention) . The application of these technologies in the industry is very mature. Many startup companies are doing similar commercial applications. For example, taking pictures for product shopping guide recommendation, the company has become well-known clothing +, good clothes, and code Long technology, etc., and in the direction of translation, the commercial network focus on cross-border e-commerce multilingual translation services. As for the direction of the intelligent question answering, Chengdu Qunei Cloud Studies has taken a leading position in the industry.
So how did the startups evaluate the use of these "black technology" apps by Xiaomi?
Photo purchase
According to the millet's official introduction, photo purchase is a technical application jointly developed by Xiaomi and Taobao Taoli. The user only needs to scan the object goods around him, and the application can recommend the relevant purchase according to the recognition result, so as to guide to the product page of "Taobao Polaroid".
From the perspective of the target product category, Palitao opened up a number of product categories such as “Snack, Beauty, Bottle, Top, and Otherâ€. However, Emma, ​​the chief operating officer of Codeon Technology, revealed that Pat Li Tao currently offers only the same search. This means that if the user submits a product image search outside the database, the system may report an error. In order to verify the authenticity of the message, Lei Feng submitted a picture that was not related to the search, and the background showed that "the system was deserted."
Although there are certain restrictions on the search of pictures, Emma gave "brave" evaluation to Xiaomi's attempts to apply new technologies:
"The domestic mobile phone manufacturers generally pay attention to hardware performance, emphasizing the performance of the screen battery network, but the exploration of the software is relatively conservative. In contrast, Xiaomi is brave enough to explore the possibility of software, so he integrated Polaroid into camera applications. We are not surprised."
translation
If everybody has the impression, WeChat has also launched a similar scanning and translation function. Wechat has adopted an effective technology solution. In terms of technology, WeChat provides technology for image content recognition and calls for proper translation capabilities. Millet uses a similar method of cooperation, but it is still unknown which translation service is used.
Lei Fengnet interviewed Wei Yongpeng, a CTO of the Commercial Web site focusing on a cross-border e-commerce multilingual translation service platform.
“A lot of translation scenarios are typical of “requirements, no markets.†In the real translation market, machine translation cannot solve problems.â€
However, Internet companies can pay attention to the translation market, and it is also a good thing for companies that do translation in their professional fields, because everyone has begun to realize the value of translation.
Quiz
In recent years, more and more educational products have begun to introduce the concept of smart answers. From the step-by-step early education machine's "where it won't be where to point" to the truly intelligent judgment, there is a world of difference in the technical requirements.
In the tech project, Xiaomi cooperates with the free problem-solving application "Learn to Domineering". Xiaomi provides image recognition to detect the content of the subject and learns to provide the question bank and answer. For this application, Lin Hui, the CEO of prospective nebula, said:
"(Actually) a variant of the camera answer, then link to the next question bank answer, and nothing more. There is no ability to judge on the application level."
In addition, Lin Hui also questioned the business logic behind this application:
"The answer to the photographing cannot be collected, and it will be ignored by the teacher's parents. Essentially, this gameplay is not different from the question bank app that is installed in the mobile phone to find the answer."
As for the functions of “scanning†applications in making business cards, scanning documents, etc., Lin Hui pointed out that the technology of print recognition and fuzzy matching is mainly used, and there is not a high technical threshold.
However, industry insiders point out that the technical threshold is only part of the core barriers to the enterprise, and the key is to find a suitable commercialization scenario for the technology. Whether technology can be converted into mature commercial applications requires a long trial process. The "black technology " millet should not only consider the technical issues, but also consider how to effectively combine the needs of hundreds of millions of users with technology products and form a business logic.
Recommended reading:
Millet black technology, "run points" first face detection algorithm in the end how?
In addition to millet, there are also domestic companies that have been on the FDDB list.
F2000A Sensor Transmitter High Accuracy Detector
Jiangmen Ever-smart Intelligent Control Instrument Co., LTD , https://www.fluhandy.com