Search

A tangible understanding of the latent design knowledge base of a technology domain, i.e., the set of technologies and related design knowledge used to solve the specific problems of a domain, and how it evolves, can guide engineering design efforts in that domain. However, methods for extracting, analyzing and understanding the structure and evolutionary trajectories of a domain’s accumulated design knowledge base are still underdeveloped. This study introduces a network-based methodology for visualizing and analyzing the structure and expansion trajectories of the design knowledge base of a given technology domain. The methodology is centered on overlaying the total technology space, represented as a network of all known technologies based on patent data, with the specific knowledge positions and estimated expansion paths of a specific domain as a subgraph of the total network. We demonstrate the methodology via a case study of hybrid electric vehicles. The methodology may help designers understand the technology evolution trajectories of their domain and suggest next design opportunities or directions.

Greenhouse gas emissions from anthropogenic sources are believed to be the main cause of global warming and climate change. Furthermore, fossil fuels are forecasted to remain the dominant source of energy in the near future. Therefore, capture and sequestration of greenhouse gases and in particular carbon dioxide is likely to be a major pathway toward environmental protection and energy sustainability. Such clarity has stimulated an intense and diverse range of research into various capture and mitigation technologies, which race with global warming in real-time. Quantification of the performance improvement rates of these technologies can inform decision-makers’ long-term investment strategies, and help researchers to identify technical bottlenecks and research potential. The present research estimates the yearly performance improvement rate of CO2 capture, non-CO2 GHG capture, and GHG mitigation technologies, using a novel method based on patent data and the corresponding citation network. Our findings suggest that membrane-based, condensation/rectification-based, and adsorption-based carbon capture are the most promising technologies. Furthermore, the average CO2 capture technologies are likely to improve slightly faster than solar, wind, and battery technologies, indicating their important role in future electrical grids. Unfortunately, subterranean or submarine CO2 storage was identified as a slow-growing technological domain and potentially a bottleneck in the future sustainable carbon chain, which requires further efforts.

This article accompanies the study presented in Triulzi et al. (2020) [1]. It briefly describes and makes available the data on functional performance for 30 technology domains, their patent sets, the measurement of patent centrality and the method to estimate the yearly technology performance improvement rate (TIR) that underly that study. Some of this data (performance time series and the lists of patents for 28 domains) has been collected by other authors for previous studies but were previously unavailable to the public. Measurements of patent centrality and other patent-based indicators for the 30 domains, and for 5.259.906 utility patents granted by the United States Patent and Trademark Office between 1976 and 2015 are novel data contributed by Triulzi et al. (2020) [1]. Here we organize, describe and make available the collection of data in its entirety. This allows anyone interested to replicate the study or use the method to estimate the improvement rate of a given technology for which patents can be identified. For a detailed description of the data and methods see Triulzi et al. (2020) [1].

In this work, we attempt to provide a comprehensive granular account of the pace of technological change. More specifically, we survey estimated yearly performance improvement rates for nearly all definable technologies for the first time. We do this by creating a correspondence of all patents within the US patent system to a set of technology domains. A technology domain is a body of patented inventions achieving the same technological function using the same knowledge and scientific principles. We obtain a set of 1757 domains using an extension of the previously defined classification overlap method (COM). These domains contain 97.14% of all patents within the entire US patent system. From the identified patent sets, we calculated the average centrality of the patents in each domain to estimate their improvement rates, following a methodology tested in prior work. The estimated improvement rates vary from a low of 1.9% per year for the Mechanical Skin treatment - Hair Removal and wrinkles domain to a high of 228.8% per year for the Network management - client-server applications domain. We developed a one-line descriptor identifying the technological function achieved and the underlying knowledge base for the largest 50, fastest 20 as well as slowest 20 of these domains, which cover more than forty percent of the patent system. In general, the rates of improvement were not a strong function of the patent set size and the fastest improving domains are predominantly software-based. We make available an online system that allows for automated searching for domains and improvement rates corresponding to any technology of interest to researchers, strategists and policy formulators.

Hydrogen economy is at a crucial point. The market demands clean and sustainable energy and fuel cell technologies look viable and quite appealing for a broad range of applications. Moreover, fuel cells are not only clean but also efficient and flexible and, among them, solid oxide fuel cells are very promising. The main problem is to understand which development stage various fuel cell technologies have reached and their yearly performance improvement rates. This information can provide insight into the barriers and the key drivers of innovation of the different types of fuel cells. Furthermore the differences in performance improvement rates could suggest the research direction that the fuel cell industry is taking. In a few words, the combination of patent analysis, bibliometrics, and rationalization of fuel cell technologies can help us to have a complete picture of their technological development. This chapter aims at providing such an overview.

In this work, we attempt to provide a comprehensive granular account of the pace of technological change. More specifically, we survey estimated yearly performance improvement rates for nearly all definable technologies for the first time. We do this by creating a correspondence of all patents within the US patent system to a set of technology domains. A technology domain is a body of patented inventions achieving the same technological function using the same knowledge and scientific principles. We obtain a set of 1757 domains using an extension of the previously defined classification overlap method (COM). These domains contain 97.14% of all patents within the entire US patent system. From the identified patent sets, we calculated the average centrality of the patents in each domain to estimate their improvement rates, following a methodology tested in prior work. The estimated improvement rates vary from a low of 1.9% per year for the Mechanical Skin treatment- Hair Removal and wrinkles domain to a high of 228.8% per year for the Network management- client-server applications domain. We developed a one-line descriptor identifying the technological function achieved and the underlying knowledge base for the largest 50, fastest 20 as well as slowest 20 of these domains, which cover more than forty percent of the patent system. In general, the rates of improvement were not a strong function of the patent set size and the fastest improving domains are predominantly software-based. We make available an online system that allows for automated searching for domains and improvement rates corresponding to any technology of interest to researchers, strategists and policy formulators.

Technology is a complex system, with technologies relating to each other in a space that can be mapped as a network. The technology network's structure can reveal properties of technologies and of human behavior, if it can be mapped accurately. Technology networks have been made from patent data, using several measures of proximity. These measures, however, are influenced by factors of the patenting system that do not reflect technologies or their proximity. We introduce a method to precisely normalize out multiple impinging factors in patent data and extract the true signal of technological proximity, by comparing the empirical proximity measures with what they would be in random situations that remove the impinging factors. With this method, we created technology networks, using data from 3.9 million patents. After normalization, different measures of proximity became more correlated with each other, approaching a single dimension of technological proximity. The normalized technology networks were sparse, with few pairs of technology domains being significantly related. The normalized network corresponded with human behavior: we analyzed the patenting histories of 2.8 million inventors and found they were more likely to invent in two different technology domains if the pair was closely related in the technology network. We also analyzed 250 thousand firms' patents and found that, in contrast, firms' inventive activities were only modestly associated with the technology network; firms' portfolios combined pairs of technology domains about twice as often as inventors. These results suggest that controlling for impinging factors provides meaningful measures of technological proximity for patent-based mapping of the technology space, and that this map can be used to aid in technology innovation planning and management., Comment: 13 pages + 23 pages Appendix and SI

The future direction of technology development depends on the relative yearly rate of functional performance improvement of different technologies. We use patent data to identify accurate and reliable predictors of this rate for 30 technologies. We illustrate how patent-based predictors should be normalized to correct for possible confounding factors introduced by changing patenting dynamics. We test the accuracy and reliability of various predictors by means of a Monte Carlo cross-validation exercise. We find that a measure of the centrality of domains’ patented inventions in the overall US patent citation network is an accurate and highly reliable predictor of improvement rates.

Data accompanying the paper "Inventors' Explorations Across Technology Domains". Code for this study is at https://doi.org/10.5281/zenodo.1035448 and https://github.com/jeffalstott/inventorexploration and as Supporting Information at the Design Science website. While that is the code and quite slim (~14MB), the data here is large (14GB zipped, ~49GB unzipped). To use, download the code from any of those sources; it will have a file structure that includes a folder data/. Then download the data here and unzip it into data/. Now you have everything you need to reproduce the study! Below is the abstract of the study: "Technologies are created through the collective efforts of individual inventors. Understanding inventors' behaviors may thus enable predicting invention, guiding design efforts or improving technology policy. We examined data from 2.8 million inventors' 3.9 million patents and found most patents are created by ``explorers'': inventors who move between different technology domains during their careers. We mapped the space of latent relatedness between technology domains and found explorers were 250 times more likely to enter technology domains that were highly related to the domains of their previous patents, compared to an unrelated domain. The great regularity of inventors' behavior enabled accurate prediction of individual inventors' future movements: a model trained on just 5 years of data predicted inventors' explorations 30 years later with a log-loss below 0.01. Inventors entering their most related domains was associated with patenting up to 40\% more in the new domain, but with reduced citations per patent. These findings may be instructive for inventors exploring design directions, and useful for organizations or governments in forecasting or directing technological change."

Technologies are created through the collective efforts of individual inventors. Understanding inventors' behaviors may thus enable predicting invention, guiding design efforts or improving technology policy. We examined data from 2.8 million inventors' 3.9 million patents and found most patents are created by "explorers": inventors who move between different technology domains during their careers. We mapped the space of latent relatedness between technology domains and found explorers were 250 times more likely to enter technology domains that were highly related to the domains of their previous patents, compared to an unrelated domain. The great regularity of inventors' behavior enabled accurate prediction of individual inventors' future movements: a model trained on just 5 years of data predicted inventors' explorations 30 years later with a log-loss below 0.01. Inventors entering their most related domains was associated with patenting up to 40% more in the new domain, but with reduced citations per patent. These findings may be instructive for inventors exploring design directions, and useful for organizations or governments in forecasting or directing technological change.

This paper applies network analysis to a citation database that combines the key references in the fields of Entrepreneurship (ENT), Innovation Studies (INN) and Science and Technology Studies (STS). We find that citations between the three fields are relatively scarce, as compared to citations within the fields. As a result of this tendency, a cluster analysis of the publications in the database yields a partition that is largely the same as the a priori division into the three fields. We take this as evidence that the three fields, although they share research topics and themes, have developed largely on their own and in relative isolation from one another. We also apply a so-called ‘main path’ analysis aimed at outlining the main research trajectories in the field. Here we find important differences between the fields. In STS, we find a cumulative trajectory that develops in a more or less linear fashion over time. In INN, we find a major shift of attention in the main trajectory, from macroeconomic issues to business-oriented research. ENT develops relatively late, and shows a trajectory that is still in its infancy.

Learning is at the base of the so-called knowledge society. New forms of learning able to more effectively challenge the complex nature of phenomenon surrounding us are increasingly necessary. In this article we argue that learning-by-modeling, through a double-loop learning process, can significantly contribute to the refinement and improvement of our knowledge of complex phenomena. To sustain this argumentation, we make use of the insights provided by an agent-based model of university-industry relationships.

In the last two decades, University-Industry Relationships have played an outstanding role in shaping innovation activities in Biotechnology and Pharmaceuticals. Despite the growing importance and the considerable scope of these relationships, there still is an intensive and open debate on their short and long term effects on the research system in life sciences. So far, the extensive literature on this topic has not been able to provide a widely accepted answer. This work introduces a new way to analyse University-Industry Relationships (UIRs) which makes use of an agent-based simulation model. With the help of simulation experiments and the comparison of different scenario results, new insights on the effects of these relationships on the innovativeness of the research system can be gained. In particular, focusing on knowledge interactions among heterogeneous actors, we show that: (i) universities tend to shift from a basic to an applied research orientation as a consequence of relationships with industry, (ii) universities' innovative capabilities benefit from industry financial resources but not so much from cognitive resources of the companies, (iii) biotech companies' innovative capabilities largely benefit from the knowledge interaction with universities and (iv) adequate policies in terms of public basic research funding can contrast the negative effects of UIRs on university research orientation.

The architecture of a firm’s network of transactions in its surrounding business ecosystem may affect its innovation performance. Here we proximate a business ecosystem as a transaction network among firms. Specifically, we analyze how the innovation performances of the firms are associated with their network positions and vertical structures in the transaction network, using the data for the Japanese electronics sector in the early 1990s. The results show that, a firm’s participation in inter-firm transaction cycles, instead of sequential transactional relationships, is positively and significantly associated with its innovation performance for vertically integrated firms. Within cycles, vertically integrated firms have better innovation performances than vertically specialized firms. Vertically integrated firms that participate in cycles have the best innovation performances in the Japanese electronics sector. These findings provide strategic implications and guidance for firms to design and manage their vertical structure and transaction network position.