Detecting technologies and skills for innovation with data-driven and Human-in-the-Loop: a Case Study on a Defene related Domain

The technological landscape is constantly evolving, leading to a rapid transformation in various industries. Emerging technologies are producing new challenges and opportunities for workers to adapt and acquire new skills [1]. Organizations are pushed to adopt different approaches for creating value and unlock the technological potential [2]. Evidence shows that the effectiveness of innovation strategies depends on the organization's culture and resource management, including engaging employees in effective use of novel practices and tools [3]. Consequently, analysis on the relationships among technological transformation and skills development are crucial to ensure workforce able to meet future demands. Such studies are a challenging task due to the dynamic and rapidly changing nature of technology and the labour market. Several authors in literature adopt qualitative and quantitative methods to analyse the impact of technological transformation on skills and the consequences on education area. Some include employer surveys [4], foresight exercises [5], expert panels [6], analytical and participatory methods [7]. The authors in [8] use Natural Language Processing (NLP) for investigating technological phenomena regarding technological emergence and convergence. NLP and the European Classification of Skills and Occupations (ESCO) is used in [9] for investigating the consequences of technological transformation on labour market. Among these possibilities, it has been highlighted that the combination of both data-driven and expert-driven approaches can lead to improved results compared to using either method in isolation [10]. Moreover, involving experts from diverse backgrounds can stimulate ideas and discussions during the sessions [11]. Therefore, studies combining different data, techniques, and expertise, are envisaged. We propose a capability-driven approach leveraging on NLP and Human-in-the-Loop, using qualitative and quantitative methods to handle the complexity of such tasks. We rely on NLP for big data analysis to delineate the current context, then we look at trends of technologies and skills in a holistic and future-oriented way, harnessing the expertise of the academic and industrial stakeholders. We focus the analysis both on technical assets (technologies) and intangible capabilities (workers and skills). This approach supports roadmap’s development forecasting the skills evolution in a technological domain. First, we collected technical and scientific documents relevant for the analysis. Then, we automatically extracted technologies, skills, applications, and job profiles, using lists and regular expressions. Next, we map their relations using job posting, industrial surveys, literature analysis, and skills and occupation taxonomy. Finally, we involved in the analysis domain-related experts to properly handle the complexity and the value of the acquired knowledge and to ensure the alignment with the needs of the sector. We will present as case study the Defence industry. The proposed method was applied in ASSETs+ (https://assets-plus.eu/), the European Blueprint for Defence, aiming to develop a strategy to up/re-skill students and professionals in Defence sector focusing on Robotics, Autonomous Systems, Artificial Intelligence, Cybersecurity, and C4ISTAR. The output is a roadmap of technologies, skills and job profiles, where technologies are positioned around the job profiles, considering the importance of a technology for a given job, and along the radial dimension, considering the level of maturity of a technology. For each job profile, a list of relevant skills is as well provided. In this way, the different areas describe the most relevant capabilities to consider in innovation strategies in the above-mentioned domains. In conclusion, the relation technology-skill is generally difficult to investigate, and even more in Defence sector, as it is at the frontier of innovation and information (both R&D and HR related) are mainly confidential for strategic purpose. The peculiarities of the sector lead us to a mixed approach: quantitative and data-driven to delineate the starting point, qualitative and holistic when looking at the future. The method can be replicated in other technological domains and in other industries. Next, the analysis provides valuable insights for stakeholders in industrial and education domain for the continuous adaptation to meet the changing demands of the workforce, as well as for policy makers to orient their actions towards the needs of the labour market.