Conference Agenda

Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).

 
Session Overview
Session
19-AM-06: ST10.2 - Emerging Technologies and New Innovation Practices
Time:
Wednesday, 19/Jun/2019:
8:30am - 10:00am

Session Chair: Jeremy Klein, RADMA
Session Chair: Jonathan Linton, University of Sheffield
Location: Room PC 21

Session Abstract

R&D and innovation are often based on generic concepts and frameworks that are agnostic to the particular technologies involved. By contrast, this track explores how R&D and associated innovation processes are closely coupled to particular technologies. With their own characteristics and internal logics, individual technologies set the context for innovation and influence many dimensions of the innovation process, for example: timescales, research methods, skills requirements, financing requirements, risks, international topologies and information flows.

A particular focus will be with emerging ‘deep tech’ or ‘frontier technologies’ such as: machine learning, AI, big data, nanotechnology, additive manufacturing and the Internet of Things. These new technologies – with their specific characteristics – are expected to impact both how innovation is organised and how it is executed: ‘innovation in innovation’.

Papers are encouraged that concentrate on one or more specific technologies, identified at a sufficient level of granularity that their distinct characteristics can be appreciated. In this respect, it may be necessary to go beyond some of the commonly used categories such as ‘nanotechnology’ to more finely grained categories such as ‘nanomachines’.

For example, the branch of AI known as ‘deep learning’ is based on the use of multi-layer neural networks, which in turn need to be trained on a large data set. Such data sets have to be either generated from scratch or obtained from existing sources. The technology developers may not have in-house access to such data. The need for training data therefore fundamentally drives the innovation relationships or collaborations necessary for the technology to be developed. However, not all AI approaches are so dependent on training data. Looking at the technology only at the aggregated level of AI could miss insights that are important to theory, policy and practice.

As well as currently emerging new technologies, the track is open to papers that explore past examples of new technologies that have shaped innovation practices.

Finally, the track will also welcome papers that conceptualise or take an overview of the relationship between new technologies, R&D and innovation.


Show help for 'Increase or decrease the abstract text size'
Presentations

How does the CVC investment affect the venture’s innovation output: empirical evidence from GEM of Chinese venture

Yangsui Lin, Wei Liu

Chongqing University, China, People's Republic of

Context

Corporate venture capital (CVC) investment has increasingly become an important source of entrepreneurial finance, which affects the ownership structure of a venture.

Literature

while scholars have traditionally focused on understanding the main motivations behind CVC activity and its impact on the investing corporate firm(Basu, Phelps, & Kotha, 2011; Dushnitsky & Lenox, 2005a, 2005b; Wadhwa & Kotha, 2006), more recently, scholars have also started to emphasize the importance of understanding the impact of CVC investment on the investee venture(Alvarez-Garrido & Dushnitsky, 2016; Chemmanur, Loutskina, & Tian, 2014; Pahnke, Katila, & Eisenhardt, 2015; Park & Steensma, 2013). In particular, these recent studies commonly show that CVC investment has a positive effect on the venture’s innovation measured in patents and other industry-specific innovation measures, such as U.S. Food and Drug Administration (FDA) product approvals(Pahnke et al., 2015) and scientific publications(Alvarez-Garrido & Dushnitsky, 2016).

Literature Gap

Whereas the positive connection between CVC investment and the venture’s innovation output is well established in the literature, we still know little about the organizational mechanisms through which this relationship unfolds within the venture.

Research Questions

1. Do CVC ownership and founder’s management control affect the venture’s R&D investment strategy?

2. Does the interrelationship of CVC ownership and founder’s management control affect the venture’s R&D investment strategy?

Methodology

This article mainly uses empirical research method to examine the effects of CVC ownership, founder incumbency and the interrelationship of them the venture’s R&D investment strategy through regression analysis of the general linear model.

Empirical Material

The GEM of the Shenzhen Stock Exchange is a platform for entrepreneurial enterprises, small and medium-sized enterprises and high-tech industrial enterprises that cannot be listed on the Main Board for the time being. The companies in the GEM have the characteristics of high growth, strong technology orientation, and rapid development. It is in line with the preference requirements of domestic mature enterprises for CVC investment. This paper selects 164 CVC-backed venture on GEM from the year of 2009 to 2017.

Results

The empirical results show that CVC ownership and the founder’s management control have played a significant positive role in the venture’s R&D investment strategy. In particular, Based on the reason of knowledge spillover and goal consistency, the interaction between CVC ownership and the founder can further promote venture’s R&D investment.

Contribution to Scholarship

Based on the theory of shareholder governance, this paper examines the impact of the proportion of CVC equity on the R&D investment of the venture. The result shows that the ratio has a significant positive effect on the innovation investment of venture rather than the inverted U-type relationship, enriching the relevant literature in the domestic CVC field from the perspective of shareholder governance. In addition, This article views from the perspective of CVC-backed venture, and focuses on the impact of the interaction between CVC and its founders on R&D investment, providing a new research perspective for the mechanism of investment in technological innovation.

Contribution to Practice

The research results suggest that CVC-backed venture could appropriately increase the ownership and the number of board of directors of CVC company in a new venture, which could encourage the desire of CVC company to provide complementary assets and promote the communication between CVC company and venture.

Fitness

This study has explored the effects of CVC ownership, founder, and the CVC investor–founder interaction on research and development (R&D) investment strategies in technology-based entrepreneurial ventures, which is relative to the emerging technologies and innovation practice.

Bibliography

Alvarez-Garrido, E., & Dushnitsky, G. (2016). Are entrepreneurial venture’s innovation rates sensitive to investor complementary assets? Comparing biotech ventures backed by corporate and independent VCs: Comparing Biotech Ventures Backed by Corporate and Independent VCs. Strategic Management Journal, 37(5), 819–834. https://doi.org/10.1002/smj.2359

Basu, S., Phelps, C., & Kotha, S. (2011). Towards understanding who makes corporate venture capital investments and why. Journal of Business Venturing, 26(2), 153–171. https://doi.org/10.1016/j.jbusvent.2009.07.001

Chemmanur, T. J., Loutskina, E., & Tian, X. (2014). Corporate Venture Capital, Value Creation, and Innovation. Review of Financial Studies, 27(8), 2434–2473. https://doi.org/10.1093/rfs/hhu033

Dushnitsky, G., & Lenox, M. J. (2005a). When do firms undertake R&D by investing in new ventures? Strategic Management Journal, 26(10), 947–965. https://doi.org/10.1002/smj.488

Dushnitsky, G., & Lenox, M. J. (2005b). When do incumbents learn from entrepreneurial ventures? Research Policy, 34(5), 615–639. https://doi.org/10.1016/j.respol.2005.01.017

Pahnke, E. C., Katila, R., & Eisenhardt, K. M. (2015). Who Takes You to the Dance? How Partners’ Institutional Logics Influence Innovation in Young Firms. Administrative Science Quarterly, 60(4), 596–633. https://doi.org/10.1177/0001839215592913

Park, H. D., & Steensma, H. K. (2013). The Selection and Nurturing Effects of Corporate Investors on New Venture Innovativeness: Role of Corporate Investors on New Venture Innovativeness. Strategic Entrepreneurship Journal, 7(4), 311–330. https://doi.org/10.1002/sej.1165

Wadhwa, A., & Kotha, S. (2006). Knowledge Creation Through External Venturing: Evidence from the Telecommunications Equipment Manufacturing Industry. Academy of Management Journal, 49(4), 819–835. https://doi.org/10.5465/amj.2006.22083132



Scientometric Mapping of Deep Reinforcement Learning using Text Mining Techniques

Sercan Ozcan1, Jbid Arsenyan2, Aldo Stornelli1, C. Okan Sakar3

1Portsmouth Business School, University of Portsmouth, United Kingdom; 2Rennes School of Business, France; 3Bahcesehir University, Turkey

Context

Global bibliometric examinations and expert views on Machine Learning (ML) show that its future resides on Deep Learning (DL), Reinforcement Learning (RL), and Deep Reinforcement Learning (DRL). This work presents a scientometric mapping of DRL studies in relation to DL and classical RL using text mining techniques.

Literature

In ML, there are three general learning approaches: supervised, unsupervised, and RL. While DL techniques were proposed a long time ago, recent advances in hardware technology and distributed computing have enabled them to be applied to these learning approaches (LeCun et al., 2015). Specifically, integration of DL with RL, referred as DRL, helped the ML community to take a crucial step towards AGI (Garnelo et al., 2016; Arel, 2012). Therefore, even though supervised and unsupervised learning have been applied to many real-life problems successfully, DRL has the highest potential to solve complex problems such as self-driving, unmanned self-operating vehicles, dialogue generation, continuous control, and enhanced image classification (Li, 2017).

Literature Gap

Klinger et al. (2018) examines the state-of-the-art of DL by mapping the development of the AI general purpose technology. There are no studies on RL nor DRL using scientometrics approach. Our study aims to fill this gap by examining these three domains, and the intersection of DL and RL.

Research Questions

What are the key cluster of technologies for DL, RL and DRL?

How DL and RL research merges to establish the future of AI?

Methodology

We categorize DL and RL publications based on their methodological and practical applications. Titles and abstracts of these studies are used to extract features. Important features are selected by using TF-IDF calculations then the co-occurrence matrix of the results is generated by using VantagePoint. Afterwards, the co-occurrence matrix is used to calculate centrality measures based on degree centrality (degree, closeness, and betweenness are calculated) using UCINet. These measures are visualised by using VosViewer software to produce heat maps. Finally, the visualised data is interpreted with the support of ML experts to fulfill the aim of the study.

Empirical Material

Considering the immense impacts of these rapidly developing ML areas, we implemented a text mining approach retrieving 6,768 RL, 11,569 DL and 442 DRL scientific papers. Accordingly, we limited all our studies to journal and conference publications with minimum one forward citation rule as a quality measure. Only the studies from 2018 until March 2019 are excluded from the citation rule to have the latest scientific and technological outputs to reflect the trends of RL and DL.

Results

One of the main practical findings of this study is that RL studies have mostly focused on the theoretical background and improvement of the method, rather than its application on real-life problems. On the other hand, with the possibility of dealing with high-dimensional action space using DL, DRL has found applications in a multitude of areas covering natural language processing, self-driving, robotics, computer vision, finance, industrial control, energy optimization, wireless sensor networks, IoT, and resource allocation in various systems. “Artificial Intelligence” appears to be one of the keywords in the representation of DRL studies and it shows the significance of AI in DRL and how it is perceived by the research community for a potential AGI. Our findings also show that the applications of DL are dominated by image and video processing related studies such as face recognition, pedestrian detection, video tracking, object recognition, image classification; whereas DRL has been rather used for control and optimization problems.

Contribution to Scholarship

Findings of this study contribute to the diffusion of innovation literature with regards to the ML application areas. Academic contributions are threefold: First, we categorize all DL, RL and DRL studies using a clustering method. Second, we offer an illustration of technological convergence into DRL and diffusion of ML approaches into different application areas. Finally, we present fusion trajectories with regards to DL and RL into DRL.

Contribution to Practice

Darwich (2018) argues that not understanding the recent developments in AI may misdirect resources. This work undertakes the task of mitigating this risk and offers a unique insight into the future of AI by mapping developments in RL, DL, and DRL domains. We illustrate key technological and scientific developments expected to impact industries through AI innovations.

Fitness

Our study investigates Deep Learning, Reinforcement Learning, and Deep Reinforcement Learning, which present the future of R&D and Innovation. The applications of Deep Reinforcement Learning fit perfectly into "Emerging technologies and new innovation practices", hence the choice of this particular track.

Bibliography

Arel, I. (2012). Deep reinforcement learning as foundation for artificial general intelligence. In Theoretical Foundations of Artificial General Intelligence (pp. 89-102). Atlantis Press, Paris.

Darwich, A., (2018).Human-level intelligence or animal-like abilities? Communications of the ACM, 61(10), 56-67.

Garnelo, M., Arulkumaran, K., & Shanahan, M. (2016). Towards deep symbolic reinforcement learning. arXiv preprint arXiv:1609.05518.

Guan, J., & Zhao, Q. (2013). The impact of university–industry collaboration networks on innovation in nanobiopharmaceuticals. Technological Forecasting and Social Change, 80(7), 1271-1286.

Huang J., (2016). Patent Portfolio analysis of the cloud computing industry. Journal of Engineering and Technology Management, 39, 45-64

Klinger, J., Mateos-Garcia, J. C., & Stathoulopoulos, K. (2018). Deep learning, deep change? Mapping the development of the Artificial Intelligence General Purpose Technology. Mapping the Development of the Artificial Intelligence General Purpose Technology.

LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521, 436-444.

Li, Y. (2017). Deep reinforcement learning: An overview. arXiv preprint arXiv:1701.07274.

Porter, A. L., & Cunningham, S. W. (2004). Tech mining: exploiting new technologies for competitive advantage (Vol. 29). John Wiley & Sons.

Rafols, I., & M. Meyer (2007), How cross-disciplinary is bionanotechnology? Explorations in the specialty of molecular motors. Scientometrics, 70 (3), 633-650.

Shapira P., Youtie J., & Kay L., (2011). National innovation systems and the globalization of nanotechnology innovation. The Journal of Technology Transfer, 36, 587-604.



Modelling the impact of policy, producers, and users on the application of the Internet of Things

Wei Xu1, Jonathan Linton1,2

1University of Sheffield, United Kingdom; 2Higher School of Economics, Moscow

Context

Applying the Internet of Things (IoT) differs from many emerging technologies due to the potential for loss of control and unanticipated usage of the information produced by the associated information provided by sensors and sensor networks. Consequently, traditional models of innovation diffusion and adoption cannot be readily applied.

Literature

The importance of policy and regulation to the adoption and extraction of value from the Internet of Things has already been established in the literature, although the technology is youthful. The existing user and producer innovation literature is integrated with policy concerns to act as a base for assessing the existing requirements, barriers and enablers relating to the integration of IoT into existing products and the development of completely new products from IoT.

Literature Gap

The interaction of policy and regulation as a driver of adoption/diffusion is considered and demonstrated through the assessment of how existing IoT

Research Questions

Hence, the question addressed are:

(1) Do policy and regulation interact with the adoption and diffusion of IoT?

(2) How do policy

Methodology

Archival analysis is conducted through the study of IoT policy and regulation that has or is in the process of being adopted. Our analysis is based on reports in the existing literature. While 80 different sources discuss the role of policy and regulation for IoT, the substantive coverage is limited to 14 separate jurisdictions. The nature of policy and regulation is mapped against produce and use innovation systems. In doing so differences in coverage and gap between the jurisdictions are identified and assessed.

Empirical Material

Archival analysis is conducted through the study of IoT policy and regulation that has or is in the process of being adopted. Our analysis is based on reports in the existing literature. While 80 different sources discuss the role of policy and regulation for IoT, the substantive coverage is limited to 14 separate jurisdictions. The nature of policy and regulation is mapped against produce and use innovation systems. In doing so differences in coverage and gap between the jurisdictions are identified and assessed.

Results

From consideration of policy and regulatory initiatives in 14 jurisdictions, we find that the Producer User Innovation framework (Linton and Fursov, 2017) is impacted by policy and regulation decisions. Failure to consider the regulatory environment can result in the model providing misleading results. Furthermore, the model can be used to assist in development of more effective policy and regulation.

Contribution to Scholarship

While policy and regulation are know to have an impact on adoption and diffusion, the interaction of these two factors has not been studied in relation to models of producer and user innovation.

Contribution to Practice

This paper offers insight into the development of appropriate policy and regulation for encouraging and directing the application of Internet of Things (IoT). Furthermore, the insights provided are applicable to other technologies that are information rich - such as cyber supply chain integration and artificial intelligence.

The development, demonstration and testing of a model that assesses the presence and absence of barriers to adoption, makes the selection of the appropriate policy and regulatory levers to build a bridge from industrial research to society achieving economic and social benefit more evident.

Fitness

This paper contributes to this year’s conference theme the innovation challenge bridging research, industry and society by offers insight into how policy and regulation can be used more effectively to satisfy the needs of society and industry when considering information-rich technologies such as Internet of Things (IoT) and Artificial Intelligence (AI).

Bibliography

Linton, J.D., Fursov, K.S., 2017 “Modelling the future structure of innovation”, Seventh International Conference on Foresight and STI Policy, National Research University Higher School of Economics, Moscow, November.



 
Contact and Legal Notice · Contact Address:
Privacy Statement · Conference: R&D Management Conference 2019
Conference Software - ConfTool Pro 2.6.128+TC
© 2001 - 2019 by Dr. H. Weinreich, Hamburg, Germany