Hi everyone, welcome to the 9th chapter in our Tencent Cloud Solutions Architect Professional course, AI Solutions. At the end of this chapter, you'll be able to understand the development and application of AI technology, Tencent Cloud's AI service system, and Tencent Cloud's AI solutions. In this chapter, we'll cover three sections : the Development and Application of AI, Tencent Cloud's AI Service System, and Tencent Cloud's AI Solutions. This video will cover the first section: the development and application of AI. Subsequent videos will cover the remaining two sections. Let's get started with section 1, the development and application of AI. In this section, we'll cover the basic concepts of AI, the evolution of AI, and the AI industry chain. Artificial intelligence, AI, is a general term for products and related services that enable machines to have capabilities such as perception, analysis, reasoning, and decision-making. AI consists of machine learning and deep learning frameworks. The basic process of machine learning is as follows : research question definition, data collection, data pre-processing, model training and tuning, model validation, model assembly, and release and run. The three elements that underpin this process are data, algorithms, and models. Essentially, data must be collected, prepared, and used for model training for machine learning until the machine is able to predict patterns in data with high accuracy. The common algorithms of machine learning include supervised learning, unsupervised learning, and reinforcement learning. For supervised learning, regression algorithms include random forest decision trees, gradient boosting decision trees, isotonic regression, and linear regression. While classification algorithms include logistic regression, naive Bayes, linear support vector machines, decision trees, random forest, GBDT, and K-nearest neighbors. For unsupervised learning, clustering algorithms such as K-means clustering, Gaussian mixture, bisecting K-means clustering, and spectral clustering may be used. Additionally, association algorithms such as FP-Growth Apriori, and dimensionality reduction algorithms such as PCA and ICA may also be utilized. For reinforcement learning, Q-learning and DQN algorithms can be used as well. The evolution of AI is shown on this timeline and is split into four stages from 1956 to the present. During stage 1 in 1956, the Dartmouth Conference marked the birth of AI, and the neural network perception was invented by Rosenblatt in 1957. During stage 2, limited by computing power, the industry experienced its first stagnation period in 1979. The Xcon expert system was released and began to see the economic benefits in 1980. During stage 3, the Xcon expert system saved millions of dollars per year for enterprises from 1990 to 1991, and the high-dimensional kernel space form of support vector machines, SVMs was proposed by Vapnik. Stage 4 starts with Hinton's proposal of a deep learning neural network in 2006. Additionally, Apple and Google implemented AI technology and voice assistance and self-driving cars in 2011 to 2012. In 2013, deep learning algorithms made a breakthrough in speech and visual recognition. Currently, we're still in stage 4 of the AI evolution. The AI industry chain consists of software, hardware, basic services, the technology layer, the application layer, and smart industries. Software and hardware consist of machine learning platforms, speech development platforms, open image platforms, and brain-inspired intelligence platforms. Basic services include natural learning processing, speech recognition, and computer vision. The technology layer consists of semantic understanding, machine translation, knowledge graphs, speech recognition, text to speech, image recognition, image processing, and OCR. While the application layer implements feature recognition, smart transportation, smart robots, and smart devices. AI industry applications include smart healthcare, smart transportation, smart manufacturing, smart finance, smart retail, smart communication, and smart education. Now let's look at AI technologies and applications. Computer vision implements machine vision in which computers replace human eyes to recognize, track, and measure targets. Meanwhile, NLP involves trying to understand the literal and context-specific meanings of words. Language technology involves the use of signal processing and recognition technologies to make Machines able to automatically recognize and understand spoken languages and convert them into text and commands. AI solves industry-specific challenges across the government, finance, and healthcare industries. For government challenges, technologies such as computer vision and machine learning may be used to increase the proportion of self-service. The analysis of suspects' life patterns and possible whereabouts can be implemented using big data and the computer vision technology can be used to find and arrest them. For finance, you can use speech recognition, semantic understanding, and other technologies to build a smart customer service to solve users' service problems and reduce customer service costs. Big Data and AI can also be used to develop smart investment advisory services and provide personalized custom service to more customers. Additionally, you can combine AI and big data to build a smart risk management system that assesses data comprehensively in multiple dimensions, improving risk management capabilities. In healthcare, smart imaging can be used to quickly screen early cancer symptoms and help patients detect lesions earlier. Furthermore, you can access health management services through various devices such as mobile devices to change people's health habits from the source. For transportation, self-driving technology can be used to liberate human hands and perception through sensing and computer vision and empower shared transport and unmanned logistics to greatly improve the efficiency of mobility and logistics. For retail, machine learning can be used to generate user profiles and place ads based on user preferences. In addition, machine vision technology can be used to understand customers' behaviors, analyze their actual needs, and improve the consumer experience with technologies such as computer vision, speech and semantic recognition, and robots. For manufacturing, computer vision technology can be efficiently and accurately used to detect defective products, and robots can replace human workers in dangerous lines of work.