PI: Christopher L. Magee, Institute for Data, Systems and Society, MIT
Abstract
Prior work on the emergence of new technologies has been hampered by lack of sufficiently detailed data on numerous enough cases to adequately test concepts and models for such emergence. Recent work in the PI’s Design and Invention Group has made significant progress in key areas that allow us to now propose moving forward with a data and modeling rich study of four important modern technology cases (MRI, Genome Sequencing, CRISPR, 3D printing). The key recent developments in the DIG group that make this possible are:
1. The conceptual base of a technological domain with functional and scientific knowledge bases;
2. A patent classification methodology – the Classification Overlap Method(COM) that enables collection of highly relevant and complete patent sets for arbitrary technological domains;
3. Path centrality analysis for all patents in the USPTO;
4. Link prediction for future technological domain interactions;
5. Scientific publication and patent domain integration.
Report
- Project Title: Technological Domain Formation
- Principal Investigator: Christopher Magee, Institute for Data, Systems and Society, MIT Engineering Systems Division, co-director of the International Design Centre, Singapore University of Technology & Design
- Grant Period: September 2017 – August 2018
Summary:
An initial exploration of the important phenomenon of domain emergence was accomplished during this MIT/Skoltech project. Significant progress was made in deepening our understanding of what happens when a new domain emerges. In particular, new domains are usually derived during much higher than normal citations of newly discovered science by the initial patents that serve as the seed for the domain.
Methods and Data:
The methods described in the proposal for this project were largely followed in the execution. Pre-existing patent sets for technological domains were analyzed to determine which involved emergence of the domain within the period for which patents can be collected- after 1976. Analysis of the domains to ensure that the domain was a new domain and not just a novel continuation of an existing domain was added to the methods described previously. Five domains then made up the studied group - 3D printing, Genome Sequencing, MRI, Optical information storage, and optical telecommunication. For each of these domains, the initial 20 patents and the patents cited by them were gathered for more detailed analysis. Centrality of the patents, scientific publication frequency, assignee type and breadth of the roots were the factors studied in depth.
Findings and Implications:
- In all 5 domains except optical telecommunication, the assignees of the seed patents were more often individuals than in patents in the later development of the domain. Similarly, in all 5 domains except optical telecommunication, the assignees of the seed patents were less often large private firms than for patents granted later in the domain evolution (universities and governments were similar in the seed vs later). The fact that Bell Labs (AT&T) was a major assignee in optical telecommunication may account for the difference in assignees in that domain. Regardless of this one domain, it appears that new domains involve a more important role for individual inventors than does the ongoing development of the domain. This appears logical since before the domain appears, there is not a mechanism to see commercial significance.
- In all five emerging domains, the patents in the seed (and roots) cited scientific publications much more often than randomly picked patents (4 to 8 standard deviations above the mean or expected value). Moreover, the citation of scientific publications fell off from the very high initial value as the domain continued to develop and evolve. These results indicate strong importance to relatively new scientific findings in the emergence of a new technological domain. Although this result is intuitively in line with prior thought, this is the strongest objective evidence for a specific role for science that has been found in technological change research and indicates an important role for “science push”.
- The centrality of the root patents (their importance in the total patent system) was sometimes higher and sometimes lower than the seed patent centrality and the seed patent centrality was sometimes higher and sometimes lower than the patents granted during the continuing evolution of the domains. This indicates that new domains are not primarily caused by “technology push” (but prior work shows the ongoing evolution of a domain is highly correlated with its average centrality).
- The breadth of root patents is generally broader than the seed patents. This result is expected and not of great significance.