What are the breakthroughs in artificial intelligence?

AI sub-areas include: Machine Learning (ML), Natural Language Processing (NLP), Deep Learning (DL), Robot Process Automation (RPA), Regression, and more. So what have the AI ??achieved in the past year? After we chatted with 21 professionals, we gathered their insights.

Fact

In the past year, artificial intelligence has made many breakthroughs, especially in terms of deep learning. For example, AlphaGo Zero is able to learn Go and Chess by itself and play games with humans without human intervention. The voices generated by Taco Tron and Baidu's DeepVoice are almost identical to human languages. In addition, computer vision, target detection and image segmentation have become more accurate, even in human medical diagnostics and biological research. However, technologies such as natural language processing, chat bots, and text summaries have not met expectations.

Artificial intelligence has been around for a long time, both old and new things are improving. It is important not to underestimate the power of public awareness. When Deep Blue defeated Gary Kasparov, the situation was different. It was only in the movie that human beings were defeated by machines, but now they have actually happened, which has greatly changed people's views. And we have a lot of applications that provide business value through artificial intelligence.

Artificial intelligence is no longer considered to exist only in science fiction. Most technology companies already understand the benefits of artificial intelligence to the enterprise. This has enabled the technology to grow rapidly over the past few months, with better profitability and the ability of machines to improve their learning processes in real time.

In the past year, we have focused on building true conversational AI. Today's various assistants do not have the ability to handle more complex and valuable tasks, and artificial intelligence is required to achieve them. It can reason based on knowledge, understand incomplete or vague language through context and personalization, and artificial intelligence utilizes and transcends pattern matching to achieve true dynamic dialogue. Just as humans communicate through gestures, gaze, and other factors, we also begin to connect other services and virtual assistants to the system. That's why we introduced the cognitive arbitrator, which seamlessly connects and integrates different virtual assistants, third-party services and content through a single interface across the automotive, smart home and Internet of Things (IoT) ecosystem to complete complexities. Tasks and enhance the user experience. Therefore, we are able to provide users with unique and individualized experiences to the maximum extent, while realizing the interactivity of various services between the assistants. This is a win-win for every individual in the IoT ecosystem, especially those who buy products and services.

AI and ML have moved out of the lab and moved to more mainstream applications. Artificial intelligence is entering the new charter and is just beginning. Six years ago, the title of data scientists did not exist, and now it has become very specialized, and data scientists and developers have realized the use of artificial intelligence to complete tasks faster and better.

GPUs

From 2000 to 2003, all trading companies gradually adopted algorithmic trading. In the past few years, machine learning has grown rapidly due to increased application requirements. In some situations where creativity is needed, artificial intelligence is replacing humans because machines can make their own decisions based on new sources of signals and large amounts of data.

Technically, GPU-based servers have become commonplace in the past year as developers have begun to take advantage of processing power to accelerate application development. Professional processors like Google's TPU are starting to emerge, and its rival cloud service providers are working together to develop an open source deep learning library. In addition, steadily transition from big data and point tools such as Hadoop and Spark to a broader range of data analysis using artificial intelligence and neural networks. ML narrows the gap between these methods by using large, different data sets and applying algorithmic intelligence to the analysis. The self-learning ability of learning algorithms is still in its infancy, the position of artificial intelligence in our lives is increasing, the product and service recommendation engine and image processing system have been significantly improved, and artificial intelligence has produced many new occupations. The pace of innovation in this area is rapidly accelerating.

Effectiveness

The concept of AI and ML is a key element of cloud computing, but it only works if the user has the data. Automated programs implemented through ML increase the productivity of enterprise employees, and as employees become more familiar with artificial intelligence tools, this level of automation will increase. In addition, the work of simplifying data integration is on the rise, especially as companies want to get more useful information from the data, and the growing focus on predictive analytics enables companies to turn real-time data into action guides.

Data

Artificial intelligence is not new, but its revival is due to the ability to process the required data as well as the speed and type of data. Information is large and messy, and artificial intelligence is needed to get useful information and data from it. The problem is that they don't have complete control over the surrounding data.

Artificial intelligence has undergone dramatic evolution in the past year for two main reasons: 1) All companies are rapidly transforming digitally. 2) The introduction speed of new business and operational data sets, and their increased demand for artificial intelligence to automate business and operational activities. The demand for artificial intelligence has evolved from "best fit" to "must have". Decision makers recognize that implementing artificial intelligence can make a business more successful, so artificial intelligence is now a key item on the agenda of each company's CIO and CFO.

Other

The promotion of all kinds of amnesty conveys a content that the trend of artificial intelligence will continue. The democratization of machine learning lies in the ability of ordinary engineers to use it. Compared to a year ago, software engineers can now make interesting MLs easier. With lower cost hardware, available data, and migration learning technology, you don't have to be a super professional doctor to become a subject matter expert who understands your data and controls the data, thereby realizing the commercial value of what you have learned.

Inflection point: People have gradually realized the seriousness of production problems, such as the shortage of data scientists. In order to solve this problem, there is now a large amount of online education, and the university has opened a data science course. So the realization of the national data scientists, and the trend of automated ML: machine automatic assisted algorithm to make choices.

They are in the absence of a cloud computing skill set and no data scientists. I have been studying how to implement smarter computing on edge devices for years. With semantic intelligence ML, edge devices can be made smarter. Can we make these edge device systems do some memory tasks? Of course, this requires more diverse device deployments and the incorporation of instantiated digital roles and applications into the component model. Thereby making the semantics richer.

Of course we will also get bored with deep learning and black box technology. On the research side, it seems that a big shift has taken place, starting to turn to algorithms that are less transparent and have less data. How can we conclude with real data without using big data? Some systems have a very large amount of data, and some are not. How do we use statistics and other data techniques to derive meaningful solutions?