1. Application of artificial intelligence in speech language industry and market scale analysis
Artificial intelligence speech language technology is an information processing technology that enables people and machines to use language as a link. Human-machine dialogue converts speech into words for machine processing through audio collection and signal processing of sound signals. After the machine performs speech recognition and semantic understanding, it carries out dialogue management, natural language generation, and converts text language into sound for output through speech synthesis technology, Finally, a complete human-computer voice language interaction is formed.
The industrial chain of artificial intelligence speech language market can be divided into six links according to key technologies, and each link can be further grouped into three modules: acoustics, speech perception and language cognition.
Most companies in the artificial intelligence speech language industry only focus on a single or part of the industrial chain, and few companies can have technologies, products and services covering all links of the industrial chain. At present, there are about 400 companies in the domestic artificial intelligence speech language industry, and only a few can achieve the coverage of the whole industrial chain.
Under the catalysis of the epidemic, intelligent applications in various industries have ushered in a demand inflection point and entered a demand explosion period. It is estimated that the total development space of consumer application scenarios will exceed 70 billion yuan in 2030. Smart home, smart driving, smart office and other enterprise level scenarios are accelerating under the catalysis of the epidemic. The market demand is constantly expanding, and the development space is expected to reach a scale of 100 billion.
Intelligent voice language technology makes human production and lifestyle gradually change. After receiving information such as the user‘s voice, human-computer interaction products based on intelligent voice language technology can convert the user‘s intention into content that the machine can understand and further process, so as to help the user solve problems or complete specific tasks. Among them, the dialogue robot can reduce the human cost, reduce the labor workload, improve the work efficiency, and solve the needs of user customer service, marketing, quality inspection, incoming and outgoing calls; Consumer grade intelligent hardware equipped with human-computer interaction functions, such as smart home appliances, smart cars, smart wearable devices, etc., can provide richer device interaction functions through voice language interaction and improve the convenience of device manipulation.
In 2021, the market scale of China‘s dialogue human-computer interaction core products will reach 9.150 billion yuan, driving the economic scale of relevant industries to 74.26 billion yuan. It is expected that the scale of core products will reach 23.7 billion yuan in 2025, driving the scale of relevant industries to 152.5 billion yuan.
2019-2026 China‘s human-computer interaction core products and drive the scale of related industries
The combination of artificial intelligence and real economy is more and more, and the in-depth combination with application scenarios will produce greater commercial value. In recent years, artificial intelligence voice language technology has been widely used in various industries, including home appliances, automobiles, consumer electronics, finance, logistics, real estate, government affairs, medical care and so on. In 2020, the scale of core products of intelligent voice language technology applied in various vertical industries will reach 5.770 billion yuan, driving the scale of relevant industries to 31.770 billion yuan. It is expected that the scale of core products will reach 15.91 billion yuan in 2025, driving the scale of relevant industries to 87.51 billion yuan.
2. Technology trend of intelligent speech language industry
In recent years, the intelligent speech language algorithm in the industry has been constantly updated and iterated, the basic performance has been continuously enhanced, and the general recognition accuracy has no longer been the core challenge of the development of the intelligent speech language industry. The speech language technology has gradually expanded from speech perception to a full link dialogue system that integrates perception, cognition and knowledge calculation.
At the key basic algorithm level, under the conditions of controllable environment and simple structured knowledge source, the performance of voice and language processing technology has been good and reached the industrialization level, but there is still a big gap with the industrial demand in the complex real environment and natural unstructured language and knowledge processing. In terms of perception technology, research in the industry has gradually turned to focus on breaking through real and complex natural scenes such as high noise, multiple interference and low resources at the end side; In terms of cognition and Knowledge Computing, we will focus on understanding knowledge Q & A, dialogue understanding and management technology, as well as deep knowledge structure in professional fields, and further enhance the intelligence oriented knowledge atlas, dialogue Q & A, reading comprehension, translation and other capabilities in vertical fields. On the other hand, the needs of personalization, scenario customization and privatization deployment have become the common needs of traditional industries for intelligent transformation and digital upgrading, such as personalized voice reproduction, question and answer dialogue in the new semantic field, and privatization identification deployment to protect privacy. The small data migration learning and autonomous learning algorithms supporting this demand and their combination in various fields of speech and language processing are also the development trend of algorithm technology in the intelligent speech and language industry.
(2.1) the emergence of full duplex voice makes human-computer interaction more natural and smooth
Full duplex is a term in the communication discipline, which means to allow data to be transmitted in two directions at the same time. It is applied in the intelligent voice language industry, that is, real-time and two-way voice information interaction, which is a conversation mode in the context of people‘s improvised free interaction. Different from single round interaction and multi round interaction, full duplex can "listen, think and speak", think while receiving voice information, and realize dynamic prediction, so as to answer at a faster speed, making human-computer interaction more natural and smooth; At the same time, full duplex voice can also achieve rhythm control. According to the importance of the user‘s answer, it can decide whether to interrupt or continue listening, whether to complete the previous question first or answer the user‘s additional questions first; In addition, full duplex voice can also understand the scene, identify whether the user is currently talking with AI, and adjust the volume and tone according to different objects and scenes. In the future, the application scenarios of intelligent voice language will become more diverse, and the environmental conditions will become more complex. The advantages of full duplex voice will become more prominent and become the mainstream interaction mode in the intelligent voice language industry.
(2.2) optimize human-computer interaction experience, and multimodal interaction has become an inevitable trend
In the process of interaction, human beings do not communicate and communicate in isolation according to the single item of sound, expression and action, but integrate vision, hearing, touch and even smell to communicate effectively. Similarly, to make the machine more realistic and "anthropomorphic", it is necessary to promote the optimization and upgrading of human-computer interaction through the combination of voice, vision, text and other information. In response to the expanding market demand for scenario based applications of human-computer interaction, multimodal and intelligent complete solutions can better cope with complex changes in different scenarios. Multimodal interaction has become an inevitable trend of industry development.
(2.3) cognition and Knowledge Computing with dialogue and interaction as the core has become an important technical trend of intelligent information services
In the context of the explosive growth of various intelligent information software and hardware, conversational language cognitive intelligence, especially the dialogue understanding and management technology, will become the key technology of the system level integration of perception and cognition, which will greatly affect the user experience. On the other hand, it will be the development direction of Intelligent Knowledge Computing Oriented to information services to carry out knowledge structure and knowledge map construction for multiple forms of original knowledge sources such as complex structured databases and various knowledge documents in the vertical field, form a controllable human-computer interactive knowledge source, support Knowledge Q & A and dialogue, and support human decision-making.
(2.4) chip R & D is becoming increasingly critical, and end side intelligence and cloud side intelligence are applied in depth in two wheel drive AI
At present, intelligent algorithms based on deep learning usually run in cloud computing centers with strong computing capabilities. Compared with cloud computing, edge computing sinks resources and services to the edge of the network, resulting in lower bandwidth occupation, lower delay, higher energy efficiency and better privacy protection. Special purpose chips are often scenario based or specific to specific functions, and their cost and efficiency are much better than general-purpose chips, which can further improve the computing efficiency of the product end side and enhance the optimization adaptability for specific application scenarios. In the future, the development of artificial intelligence speech chips will further promote the commercialization of intelligent speech language products in various vertical industries.
|
Disclaimer: This article is transferred from other platforms and does not represent the views and positions of this site. If there is any infringement or objection, please contact us to delete it. thank you!
中恒科技ChipHomeTek
|
Service mailbox
13823783658
ChatNow