The wave of intelligence is sweeping, and the development of artificial intelligence APIs is promising

The outline of the 14th Five Year Plan points out that it is necessary to "promote the deep integration of Internet, big data, artificial intelligence and other industries, promote the development of advanced manufacturing clusters", "systematically layout new infrastructure, and accelerate the construction of the fifth generation mobile communication, industrial Internet, big data center, etc.". Under the guidance of relevant policies, the research and development of artificial intelligence technology has accelerated, and the business formats are becoming increasingly diverse. Artificial intelligence APIs, platforms, and open source communities have also become hot topics for industrial development.

On October 23rd of this year, the 2020 iFlytek 1024 Developer Festival kicked off in Hefei with the theme of "A.I. Renewal and Better". In the first half of the conference, the Senior Vice President and President of the Consumer Business Group of iFlytek released the new 1024 Plan 4.0, and also released the 2020 China Artificial Intelligence API Economy White Paper jointly released by iFlytek's development platform and iResearch Consulting.

The white paper shows that the market size of China's artificial intelligence open platform was 10.4 billion in 2019, and it is expected to reach 22.2 billion in 2020. A. I. Open platforms are still in their early stages and will continue to grow rapidly in the future. iResearch predicts a CAGR of around 41% over the next five years.

At present, the product applications developed by developers cover the Internet, education and training, smart city, robotics, finance, industry 4.0, government affairs, retail, home furnishing, smart agriculture, security and other industries. Artificial intelligence has become the key technical force for enterprise transformation.

So, what is an API? API was originally a modular architecture implemented by a single application or enterprise information system, used as an interface for communication between modules; Now it gradually evolves into an enterprise application communication mechanism and Internet cloud service communication mechanism based on REST protocol and through standard XML or JSON message format to realize service reusability.

There are many forms of APIs for artificial intelligence and machine learning. One of them is a machine learning algorithm framework that provides APIs for basic functions such as model construction, training, parameter tuning, and algorithm iteration. Popular frameworks include TensorFlow, PyTorch, and Caffe for deep learning and neural network model development, Python machine learning API libraries such as NumPy, Pandas, and Scikit learn, and distributed machine learning algorithm API libraries such as Spark MLLib.

Another approach is to publish general artificial intelligence algorithms in the form of API services, providing various types of client usage, which are typically accessed and called in a REST manner. From the current perspective, common artificial intelligence algorithms include NLP natural language processing, speech synthesis and recognition, image and video processing and recognition, etc.

There are many examples of artificial intelligence APIs, let's take a look at Google Prediction. It is reported that Google Prediction supports cloud based machine learning features, including recommendation engines, pattern recognition, prediction, and natural language processing. Developers can use this API to create applications for categorizing documents, predicting purchases, analyzing emotions, detecting spam, and more.

Microsoft has also launched three commonly used Cognitive Interaction services, including Face API, Computer Vision API, and Content Moderator API. By using the Face API, developers can obtain attribute information such as gender, age, facial features, and head posture, and now they can also obtain emotional expressions.

In fact, in addition to the booming API development, the phenomenon of accelerating the construction of open source communities for artificial intelligence is also worth paying attention to. The Chinese AI open source community provides high-quality, more open, and friendly open source platform services. The Report points out that open source platforms, including Baidu Apollo and PaddlePaddle, have a number of patents, accelerating the application and promotion of AI technology. Compared with some open source platforms abroad, Chinese artificial intelligence platforms adhere to the spirit of open source and have applied for many patents around related technologies, providing high-quality services and platforms.

The wave of intelligence in human society is approaching, but problems such as imbalanced development of the artificial intelligence industry, intensified value system conflicts, and urgent transformation of employment structure have accelerated. Next, people from all walks of life should embrace the intelligent transformation of the economy and society with a positive attitude, and promote the research and industrial development of artificial intelligence technology.

Prev: Artificial intelligence natural language processing technology drives industrial upgrading engines