R & D design is a creative intellectual activity, and also a process of synthesis, decision-making, iteration and optimization. Artificial intelligence will define its own "position" in innovative design by virtue of its advantages.
1. Build a new organizational structure and operation mode from the perspective of R & D system
Artificial intelligence makes business process automation, knowledge management automation, and management itself intelligent, including data collection, feedback, monitoring, evaluation, intelligent analysis and prediction, and intelligent decision-making. First, the allocation mode and organizational process of innovative resources are changing from producer centered to consumer centered. It has become a new driving force for enterprises to build an innovation system of deep mining of customer needs, real-time perception, rapid response and timely satisfaction. For example, in order to establish an efficient R & D Organization for the market and customers and ensure the smooth implementation of the R & D process, Chang'an automobile has carried out gradual reform of the R & D organization. From weak matrix organization to strong matrix structure, a "one vertical and two horizontal" organizational structure has been formed, which focuses on professional ability improvement vertically and product development and common basic technology horizontally, so as to realize the simultaneous promotion of multi project R & D, professional technology research and common basic research and fully provide R & D efficiency. Second, R & D mode will change from closed innovation to open collaborative innovation system. Through the establishment of a global distributed online collaborative R & D platform and mechanism, the enterprise internal collaboration and supplier external collaboration based on the unique data source are realized. For example, Haier's hope platform, the world's largest open innovation ecosystem and whole process innovation interactive community, integrates suppliers into modular suppliers, realizes modularization of product production, and enables modular suppliers to reach user needs directly, and conducts corresponding design adjustments and product research and development, upgrades modules, and builds new innovation models.
2. From the aspect of R & D and design ability, realize the efficient and lean R & D and design of enterprises
The data processing ability of artificial intelligence will gradually replace the flow part of R & D design, and bring "technical" innovation to R & D organization, which makes the traditional continuous variable design and mixed discrete variable design mode change to random variable and fuzzy variable optimization design mode. The first is to make deep use of digital technology and optimize research and development methods. Deep use of fuzzy mathematics and other theories can simplify the imprecise experience data and massive measured data in mechanical design. For example, in the process of new product design and development, digital technology is fully used, secondary development of digital tool software is used to standardize the design process, better integrate design methods, standards and specifications into digital tool software, and precipitate Enterprise knowledge. At the same time, more and more enterprises will simulate the performance of products through digital games, and constantly adjust and optimize the final product design. For example, the intelligent driving team of Stanford University developed the three-dimensional game driver seat to help in-depth neural network training by showing the formation environment of intelligent driving cars and obtaining the training data of game players. Second, R&D visualization and independent optimization to improve R&D performance. By using heuristic algorithm, genetic algorithm, ant colony algorithm and other technologies, the performance simulation, motion analysis, function simulation and evaluation of products in the design stage can be realized, which can meet the requirements of product design automation and intelligence to the greatest extent. For example, Chang'an Automobile establishes a high-performance computing system. Through independent optimization, the computing capacity of Chang'an high-performance computing platform is up to more than 10 trillion times per second. The imported R & D visualization can not only independently optimize and timely track the R & D design progress, understand the project risks, realize the real-time view of the R & D model visualization model, and ensure the high coupling cooperation among departments, disciplines and regions Same work.
3. Provide more efficient and diverse personalized customization in terms of R & D service forms
Artificial intelligence makes business process automation, knowledge management automation, and management itself intelligent, including data collection, feedback, monitoring, evaluation, intelligent analysis and prediction, and intelligent decision-making. First, relying on the big data, cloud service platform and other elements in the infrastructure layer to improve the R&D efficiency. The powerful machine learning algorithm is used to deal with the huge database, and the relevant statistics can replace the flow part of the R & D design link to improve the R & D efficiency. For example, the well-known Moore eye hospital in the UK has cooperated with Google to build a machine learning system, which can identify potential eye disease risks based on the results of digital eye scanning. Eye scanning technology has existed for a long time, but traditional machines can't analyze eye data quickly after scanning. Machine learning is an expert to deal with this kind of data, which can greatly shorten the analysis time and improve the accuracy, and a faster and more accurate outline scheme. Second, artificial intelligence will present a more intuitive and vivid form of research and development, breaking through the "impossible link" of the past. For example, under the modern science and technology, we have continuously obtained gene biology data, but it is difficult for human beings to crack and control these massive data, and can't "understand" genes. However, the in-depth learning of artificial intelligence will develop its strengths and find genes that human beings can't find Relation. When artificial intelligence finds out the healthy gene sequence pattern, human can diagnose or even predict the disease through gene, and optimize the drug target. For example, when the woodpecker company is developing new dental cleaning tools, it adopts a / B test scheme, introduces the concept of user exploration, and allows users to choose independently