Using Quantitive Ecosytem Models and Science Fiction to Prevent Stategic Technical Surprise
Abstract
The proposed work plans to create machine learning text mining information retrieval methods to identify the conceptualization and contextual scenario usages of speculative technology as introduced in science fiction novels. Additionally proposed is the creation of novel quantitative ecology based technology performance evolution methods. Combined with an empirical study to understand science fiction as a technology forecaster, these new methods create currently unavailable technology forecasting capabilities to prevent strategic technological surprise. Speculative technologies conceptualized in science fiction have shown the capability to be a technology pull through creating technologies and their usage scenarios that are subsequently realized in practice. Speculative technologies conceptualized in science fiction will be studied to provide answers to questions such as what percentage of speculative technologies are realized, what the time lag is between technology speculation and realization, what types of technologies are realized and what types are not, and importantly, which technologies are disruptive when realized? This research will approach technology interaction and subsequent performance evolution as analogous to a natural ecosystem. Recognizing that technologies evolve in an ecosystem allows access to concepts from quantitative ecology to better model and understand technology performance evolution. The work will build technology evolution models from Lotka-Volterra population dynamic models. Lotka Volterra networks of non-linear coupled differential equations have been shown to model technology performance evolution in a systems context and with better accuracy than Moore s Law. The activity proposed here is to identify realized system technologies from the science fiction analysis, identify component technology ecosystems for that system technology, and then build quantitative Lotka-Volterra systems that effectively mimic the development and realization for that technology. If successful, the proposed work will complement intelligence analysts and technology developers that actively perform technology forecasting to guide their own technology development and predict the technology development of others. This work is innovative through its formal recognition of technology ecosystems in quantitative technology performance prediction. Prior methods (such as Moore s Law) only consider single elements in building a technology forecasting model. Science fiction has been used to assist in technology forecasting; however, using science fiction to create possible technology development ecosystems has not been done. Also, emerging natural language processing and machine learning methods offer a chance to create potential use scenarios that depend on future technology capabilities.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 25, 2019
- Source ID
- HR00111910006
Entities
People
- Daniel Mcadams
Organizations
- Defense Advanced Research Projects Agency
- Texas Engineering Experiment Station