The definition of artificial intelligence technology is very broad. As an important branch of artificial intelligence research, machine learning technology has been widely applied in the fields of image and speech processing.
Software: voice, image, sensory recognition technology, software, natural language processing, advanced algorithms, data training; Hardware: deep learning chips, sensors, ICT, IOT; Big data: data collection, data storage, data calculation, data visualization;
According to the analysis of application scenarios, artificial intelligence technology can be classified as follows:
Analysis of mobile Application Performance At present, the explosive growth of mobile application software has brought a wide variety of mobile applications with a sharp increase in the number of users in the process of choosing mobile applications there is no reference standard. In this case, the friendly analysis and ranking of mobile apps can guide users to choose mobile apps, promote the healthy development of mobile apps, and improve the overall quality of mobile apps. At present, the performance analysis of mobile applications mainly focuses on end-to-end QoE analysis, user stickiness analysis, business collaboration and friendliness analysis, etc. The common point of these analyses is that they are all based on big data analysis, collecting and calculating a large amount of user and application related information and drawing conclusions.
Therefore, in the process of big data analysis and calculation, artificial intelligence algorithms are usually used to solve the problem. Different machine learning algorithms are powerful tools for big data analysis. Currently, reinforcement learning algorithm and deep learning algorithm based on neural network are widely used. The operation mode of these learning algorithms is to collect and store the energy consumption, delay, traffic, and rate of different applications in different application scenarios according to the requirements of indicators, and serve as the input end of the neural network system. Then, according to the different application types to distinguish between social applications, instant messaging, audio and video applications, cloud application, the browser application and game applications, will be different in the field of application for energy consumption, time delay and flow rate, rate of indicators such as demand analysis, and generate the corresponding index of the weighted value as the weights of the neural circuits. Finally, a neural network model is formed through training and feedback calculation of a large amount of data. The learned model can output the overall friendliness of applications, ranking applications based on current data, and predict application performance based on existing data. At present, the operators actively establish mobile Internet user behavior analysis system, on the basis of data sharing for mobile users to access the Internet behavior is analyzed, finally grasp the user’s online habits and preferences, so as to accurately locate the user demand for mobile Internet, to provide data support for mobile business analysis and decision.
Mobile Application Authentication Identity authentication is another important application direction of artificial intelligence algorithm. The most commonly used authentication methods have been the password mode of user name and password and devices such as U shield for peripherals. However, due to the risk of password being deciphered and the possibility of being forgotten, and the inconvenient problem of peripheral devices such as U shield, multi-factor authentication and biometric authentication are being widely developed. The multi-factor authentication method combines two or more authentication methods to make up for the shortcomings of different authentication methods. At present, there is a multi-factor authentication that combines password with user big data graph analysis to enhance the security of the original password login mode. The user big data atlas is a good assistant to the common password authentication mode, which can predict user behavior and reject login or payment requests that do not conform to the user behavior model by analyzing the data recorded by various behavior patterns of users on the Internet based on machine learning algorithm to complete user basic behavior model. Biometrics authentication is an authentication method that uses the unique characteristics of the user’s biometrics such as fingerprint, face, iris, finger vein and so on for comparison and identification. Biometric authentication has high security, and with the continuous development of hardware, biometric modules mounted on mobile terminals are becoming smaller and smaller, and more and more convenient to use. Due to the huge amount of biological information collection and the huge comparison database formed after feature extraction, the algorithm basis of biometric authentication relies on artificial intelligence algorithm for pattern recognition and comparison authentication, and the identification results will be analyzed and output at last.
New Application development
Ai technology also promotes new mobile Internet applications and industries. Virtual reality/augmented reality technology is based on the application of artificial intelligence technology new expansion. Virtual reality/augmented reality needs to collect the perceptual data around the user and quickly upload it to the server, and then send the result to the user’s eyewear device through server calculation. For the sake of user experience, Google provides a delay threshold of 20ms from data collection to result presentation. In other words, in order to achieve fast data exchange and computation, the transmission bandwidth is indispensable as well as the data calculation algorithm. Artificial intelligence algorithm realizes the instantaneous calculation of large amount of data, which solves the development foundation of virtual reality/augmented reality. In addition, a number of applications based on ARTIFICIAL intelligence (self-driving cars, smart homes, intelligent voice search, etc.) are developing rapidly. Baidu relies on artificial intelligence algorithms such as deep messaging network, convolutional neural network and recursive neural network to solve natural speech processing, intelligent speech recognition and search, image search and other applications. Ai technology can have a huge impact on almost every field of mobile Internet.
Free access to artificial intelligence learning materials
