The last decades witnessed a never-ending debate about the generation of innovations under different circumstances and, to a lesser extent, on adoption. With a dedicated focus on technologies, the topic of adoption and diffusion emerged only a little. However, this focus is on technologies, expressed in Technology Acceptance Models and Technology Diffusion Models, which cover only parts of the overall spectrum of innovation types. Innovation Diffusion Models similarly highlight innovation adoption but pay less attention to implementation. The challenge, however, lies more with the implementation of innovation than with initial adoption. This is especially the case for organizational innovation, which causes considerable change inside organizations. Changes inside organizations aren’t always welcome and supported by people, which resulted in the change management paradigm. The working paper highlights these general thoughts. It comes in two parts. The first focuses on Organizations, Ecosystems, and Institutional Power, and the second on Saturation and the Limits of Adoption.

Recent research applies patent autoclassification using machine learning to map the technological landscape of an industry value chain. However, when these methods are applied to emerging industries, the available patent sample data are small-scale and unevenly distributed, which cause overfitting and reduce the accuracy of patent classification. Therefore, this article proposes a framework to map the technological landscape of an emerging industry value chain through patent analysis with deep learning, which integrates a generative adversarial network as a data-augmentation method to overcome the problem of low-quality emerging-industry patent samples, and a deep neural network as a patent classifier. Based on this framework, this article conducts an application case of the 3-D printing industry. The evaluation results show that the integrated framework can effectively classify the patents with small-scale and unevenly distributed sample data, and depict the technological landscape of an emerging industry value chain. This article develops an efficient, reliable framework for patent autoclassification of emerging industries to overcome the lack of high-quality training samples, and it sheds light on the emerging industry value chain analysis with deep learning.

[目的 /意义]由于新兴技术本身的超前性,其刚出现的关注度往往不是很高.目前研究更多遵循技术发展路径依赖进行新兴技术的识别,会忽略一些颠覆现有技术轨道的技术研发.通过对与领域内主流技术相似度较低的离群专利进行分析,可以更有效地识别这类技术研发并预测新兴技术.[方法/过程]提出一种基于深度学习的离群专利识别与新兴技术预测方法.首先使用BERT预训练模型基于专利文本构建相似度网络,识别离群专利,然后基于DNN模型构建离群专利指标与技术影响力之间的关系,实现从海量离群专利中快速、准确地预测新兴技术.最后以数控系统领域为例,从德温特专利数据库获取近10年领域内所有专利,进行实证分析.[结果/结论]数控系统领域的实证分析结果验证了模型的有效性,同时对国家的技术发展政策制定以及相关领域企业技术布局具有重要的指导意义.

Radical novelty is one of the key characteristics of emerging technologies. This characteristics makes emerging technologies as a quite different from established technologies. From the perspective of radical novelty, some studies consider patents with little similarity in terms of key concepts and contents to existing patents as candidate emerging technologies. However, existing research remains in examining small-scale patents for evaluating candidate emerging technologies due to the lack of data-processing capacity-the recent rising of deep learning methods may help in this. This study, therefore, develops a novel deep learning based framework for identifying emerging technologies by combining a technological impact evaluation using patents and a social impact evaluation using website articles. Using a large scale multi-source dataset including 129,694 patents and 35,940 website articles, this paper applies the framework to investigate the case of computerized numerical control machine tool technology, through which the framework is validated. The results show that 16,131 patents out of 129,694 patents are considered as candidate emerging technologies, and 192 patents out of 16,131 patents are identified as emerging technologies through the evaluation of technology impact and social impact. This implies that these candidate emerging technologies can evolve to emerging technologies, though not all of them-we need deep learning method to scrutinize a larger scale multi-source data to identify rather a small number of potential emerging technologies. The proposed framework can also be extended to explore other disciplinary multi-source data for strategic decision support in identifying emerging technologies.

Deep learning can be used to forecast emerging technologies based on patent data. However, it requires a large amount of labeled patent data as a training set, which is difficult to obtain due to various constraints. This study proposes a novel approach that integrates data augmentation and deep learning methods, which overcome the problem of lacking training samples when applying deep learning to forecast emerging technologies. First, a sample data set was constructed using Gartner’s hype cycle and multiple patent features. Second, a generative adversarial network was used to generate many synthetic samples (data augmentation) to expand the scale of the sample data set. Finally, a deep neural network classifier was trained with the augmented data set to forecast emerging technologies, and it could predict up to 77% of the emerging technologies in a given year with high precision. This approach was used to forecast emerging technologies in Gartner’s hype cycles for 2017 based on patent data from 2000 to 2016. Four out of six of the emerging technologies were forecasted correctly, showing the accuracy and precision of the proposed approach. This approach enables deep learning to forecast emerging technologies with limited training samples.

Previous studies on identifying technology evolution pathways have ignored the point in time when an evolution pathway changes. To fill this gap in the literature, we use a novel approach based on text-mining to identify technology evolution pathways. First, the topic model was applied to discover technology topics, and topic variation through time was modeled. Second, critical junctures were detected based on topic variation to generate time segments. Finally, the main topics' changes at different segments were analyzed to identify the pathway, and a visualization of the main topics' trends were presented. To demonstrate the effectiveness of the textmining approach, we examined 3D printing technologies using 34,090 patents from 1990 to 2017, and the database we used are updated weekly. We found that traditional technologies showed a declining trend, and their practical application technologies related to products is promising. The new approach can be applied to identify technology evolution pathways characterized by critical junctures. These critical junctures are helpful in understanding technological development more clearly, especially in identifying what and when technological changes occur.

It is essential to utilize deep-learning algorithms based on big data for the implementation of the new generation of artificial intelligence. Effective utilization of deep learning relies considerably on the number of labeled samples, which restricts the application of deep learning in an environment with a small sample size. In this paper, we propose an approach based on a generative adversarial network (GAN) combined with a deep neural network (DNN). First, the original samples were divided into a training set and a test set. The GAN was trained with the training set to generate synthetic sample data, which enlarged the training set. Next, the DNN classifier was trained with the synthetic samples. Finally, the classifier was tested with the test set, and the effectiveness of the approach for multi-classification with a small sample size was validated by the indicators. As an empirical case, the approach was then applied to identify the stages of cancers with a small labeled sample size. The experimental results verified that the proposed approach achieved a greater accuracy than traditional methods. This research was an attempt to transform the classical statistical machine-learning classification method based on original samples into a deep-learning classification method based on data augmentation. The use of this approach will contribute to an expansion of application scenarios for the new generation of artificial intelligence based on deep learning, and to an increase in application effectiveness. This research is also expected to contribute to the comprehensive promotion of new-generation artificial intelligence.

The convergence of multi-disciplinary knowledge may spur emerging technologies. It is important to understand this convergence process that helps to identify these emergent technologies; however, relevant research remains sparse. Therefore, this study aims to develop a novel framework to reveal the convergence process of scientific knowledge. This novel framework integrates the machine-learning topology clustering and visualization methods, and analyzes paper citation networks. This study selects the biological–informatics domain (bioinformatics) to conduct the empirical analysis. This paper finds two major stages throughout the convergence process: the fast-changing incubation stage and the stabilized development stage. In the incubation stage, the interactions between the biology and informatics knowledge domains becomes increasingly intensive, while emergent technology is yet to form; in the stable development stage, the emergent technology starts to form as a core cluster, and based on which it grows amid stabilized knowledge interactions between the original two domains. The revelation of this convergence process contributes to the formation theory of emerging technologies that are inter-disciplinary, and is of great interest to researchers, policy makers, and industrialists.

[目的/意义]技术融合是新兴产业形成和发展的驱动力,而知识融合是技术融合的前提,对知识融合过程进行研究,对于引导新兴产业形成和发展具有重要意义.[方法/过程]首先根据现有的研究构建一个采用论文引用网络表征知识融合过程的理论模型,其次根据知识融合过程每个阶段论文引用网络的特征设计验证方法,最后以融合新兴生物信息领域为例进行实证分析.[结果/结论]生物信息领域的实证分析结果表明理论模型的有效性,可为研究知识融合过程提供一种新的方法.