Intelligent multi-layer approach for the early detection of emerging science and technology
Project leader: Andrea Schiffauerova, Concordia University
Partners: Ashkan Ebadi (National Research Council Canada)
Starting and end dates:
Starting date: September 2022
End date: March 2024
4POINT0 ecosystem(s): Artificial intelligence ecosystem
Emerging technologies can have major economic and strategic impacts. They have the potential to alter the technological paradigms on which traditional industries rely, to generate completely new industries (Porter et al. 2002; Day and Schoemaker 2000), or to modify existing socio-economic structures and production practices (Adner and Snow 2010; Rotolo, Hicks, and Martin 2015; Zhou et al. 2021). Early and accurate detection of emerging technologies can provide decision-makers coming from various R&D departments, institutions, policy-making organizations or innovation administrations with valuable knowledge, intelligence, and opportunities (Zhou et al. 2021; Jang, Park, and Seol 2021). Yet, early, identification of emerging technologies remains challenging.
Several approaches for detecting emerging science and technologies have been proposed so far. Some of them are based on machine learning algorithms (Xu et al. 2019), some on citation-based approaches (Boyack et al. 2014), and some on lexical-based approaches (Weismayer and Pezenka 2017); in addition to hybrid methods combining the above mentioned approaches (Lee et al. 2018). However, the indicators proposed by various studies are mostly descriptive and often focused on a single aspect of the problem while neglecting to consider and combine multiple emergence criteria. As such they cannot be considered as comprehensive and timely indicators of emerging technologies and are therefore not suitable to be implemented and used as the main source of insight in the decision-making processes. Moreover, the existing research studies tackling the identification of emerging technologies usually merely back-test the indicators on the historical data, and do not provide a concrete analysis and comprehensive methodology for identification of emergence in future.
This research project proposes to develop an approach which would overcome all these limitations. We propose to consider multiple parameters and attributes of emergence such as novelty, growth, and community, while using multiple approaches such as bibliometrics, social network analysis (SNA), natural language processing (NLP), machine learning (ML), and deep learning (DL) in order to perform comprehensive examination and develop an intelligent multi-layer engine incorporating numeric and text-based indicators at various levels (thematic level, field level, term level). The engine will identify future signs from scientific publications (and other sources), suggest the emergence of future science and technology and define the time window of early identification (showing how long it will take before the science or technology will in fact emerge).
It is expected that the proposed multi-aspect framework will help strategic planners, decision makers and domain experts better detect and monitor emerging science and technology trends, and thereby enabling them to identify opportunities and risks quickly and to react to them accordingly by formulating appropriate research, development, and innovation strategies.
The proposed research poses the following research questions:
- Can we develop a comprehensive multi-dimensional approach for an early detection of emerging science and technology?
- Will the developed emergence indicators be accurate and reliable, yet practical to use?
- How early before the technology gets shaped can we identify its emergence?
Ce contenu a été mis à jour le 2024-01-30 à 19 h 00 min.