According to foreign media reports, what are the hot trends of artificial intelligence (AI) this year? Let's take a look at the conclusions of industry analysts.
Trend 1: Semi-supervised AILeading practitioners in the AI ​​field believe that supervised learning is the holy grail of the AI ​​field. What is unsupervised learning? For example, a machine can “learn†what is spam without having to look at many email examples with “spam†or “non-spam†tags.
Semi-supervised learning is a transition to achieve unsupervised learning objectives, between unsupervised learning and supervised learning, that is, labeling some data, but let the machine rely on correlation analysis to guess the labels of the rest of the data. . Google has developed a technology that uses this semi-supervised learning model called graph-based learning. Using knowledge mapping techniques that perform correlation analysis among words, Google is able to eliminate the tedious task of labeling all data based on those associations. The company is already applying this technology to its many products, such as answering questions, reminders, visual object recognition, and dialogue to understand black smart email responses. Semi-supervised learning is expected to be increasingly applied to large data sets, because data labeling is a big problem for large data volumes, especially for visual and linguistic data.
Trend 2: Home voice assistants quickly spreadAccording to VoiceLabs, by the end of 2017, the number of devices supporting voice control in the United States will reach 33 million. Amazon (Alexa), Microsoft (Cortana), Google (Google Assistant) and Apple (Siri) are investing heavily in this area to create ways to attract consumers to the ecosystem of their devices. One of the ways to win consumers' favor will be to offer unique features or to offer specific discounts (including content channel subscriptions for a specific time period).
Trend 3: Social Chat RobotWeChat and other popular social media communication services in China have been promoting chat bots to help users with their daily tasks. Facebook has also just begun to integrate chat bots into its social networking platforms, including ads using linked chat bots and sponsorship ads in Facebook Messenger. These virtual assistants will become more and more popular and more and more popular. They will simplify a variety of e-commerce activities, such as booking flights and hotels, and supporting the ordering of products in the app by talking directly to the chat bot.
However, chat bots are also rapidly expanding from consumer applications to the enterprise market, helping business users. A survey of corporate executives found that 32% of speech recognition chat bots are the most commonly used type of AI technology in their workplace. Market research firm Gartner predicts that by 2020, chat bots will drive 85% of customer service interactions. Enterprise messaging and writing platforms such as Slack, Skype, Oracle, and Salesforce, as well as a large number of startups, offer software-as-a-service capabilities to help employees improve their productivity. Like smartphones, enterprise users of virtual assistants will eventually want these artificial intelligence technologies to be portable and ready to use – possibly to bring their own robotic (BYOR) movements.
Trend 4: AI as an extension of enterprise ITThe most current enterprise-level use cases that attract the most investment include: automated customer service assistants, quality management survey and recommendation systems, diagnostic and treatment systems, and fraud analysis and investigation. The enterprise-level use cases with the fastest growth in revenue in the next five years include: public safety and emergency response, pharmaceutical research and discovery, diagnostic and treatment systems, supply and logistics, quality management survey and recommendation systems, and fleet management. AI applications can use algorithms and rule-based logical reasoning to identify and respond to data flows, thereby automating a wide variety of functions across a wide range of industries and enabling employees to become more productive.
Trend 5: Development of driverless carsAccording to McKinsey & Co., unmanned vehicles are estimated to save 300,000 lives every decade because they can reduce fatal traffic accidents. This is also expected to save up to $190 billion in emergency care and injury classification expenses each year. At present, Google’s driverless car road test mileage will exceed 1 million miles, but the future focus will shift from the potential benefits of driverless cars to the necessary regulatory requirements. Legislators and policy makers will open the long process of designing and implementing new regulations. 2020 may be the first year for driverless cars to enter the market, and all sectors of society must begin to prepare for the arrival of this day. In the future, there will be more lobbying organizations in Washington, and more alliances of suppliers and users will be established, all of which lay the foundation for the widespread adoption of driverless cars.
Trend 6: Multiple alternative hardware platformsAlternative hardware platforms such as field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), and dedicated processor architectures will compete more with graphics processors (GPUs) for attention and investment. GPUs are currently the mainstream hardware platform for AI applications, especially deep learning systems. Given that the AI ​​algorithm supports changes such as automated driving or personalized medical applications with dynamic input, the processor itself has reason to be equipped with memory storage. Algorithms and workload change properties will determine which architecture is best for which application.
Trend 7: The rise of the AI ​​service marketJust as big data and data science have seen in recent years, AI-related services have also seen business opportunities, including supplier selection, implementation, training, application and algorithm development and integration, consulting and so on. As the skills and experience associated with machine learning and AI are in short supply, the scale of on-demand services offered by cloud providers is expected to expand.
Provide professional solutions for live
broadcast scenarios:
Mobile phone live broadcast scene: When
using a mobile phone to broadcast live on facebook, youtube, instagram, tiktok,
etc., the screen of the mobile phone can be wired/wirelessly projected to a
large display screen such as 32-inch/43-inch/49-inch/55-inch/65-inch, Solve the
problem that the anchor is difficult to watch the fan interaction barrage, so
that the anchor can clearly see the fan message during the live broadcast even
a few meters away from the screen.
Large-screen live broadcast scene: Install
Android system on the 32-inch/43-inch/49-inch/55-inch/65-inch large display
screen, and install the live broadcast app at the same time. Large-screen live
broadcast can enlarge the live screen.
Broadcasting studio scene: The live
broadcast large screen has a built-in studio switcher, switcher monitor and
Touch Screen.which can realize a rich live scene with multiple cameras,
multiple scenes, and multiple background switching.
Live Broadcast Display Screen,Live Streaming touch Screen,live streaming studio,Wireless Projection Screen,switcher monitor
Jumei Video(Shenzhen)Co.,Ltd , https://www.jmsxdisplay.com