Open Collaborative Data as an Innovation Platform for Machine Learning Applications
|Coordinator||Lunds universitet - Datavetenskap|
|Funding from Vinnova||SEK 495 780|
|Project duration||November 2018 - June 2019|
|Venture||Banbrytande idéer inom industriell utveckling|
|Call||Banbrytande idéer inom industriell utveckling 2018|
Purpose and goal
The project aims to explore the potential of companies sharing data with others for machine learning, similar to sharing open source software. The cost of collecting, quality assuring and maintaining data is expected to rise. For data that does not contribute to competitive advantages, we argue that open data collaboration is a way to save costs. These resources can be used to develop competitive and innovative aspects of company products and services, similar to the way open source software has revolutionized e.g. the mobile domain.
Expected results and effects
We expect the development resources for machine learning applications be utilized more efficiently when making company’s data open. This is achieved by sharing the burden of maintaining and assuring the quality of data, which many companies need but nobody can capitalize on. Furthermore, innovation may arise in the open collaboration between companies on topics beyond what they are directly collaborating on. Providing means to collaborate, not only on OSS and OI but also on Open Collaborative Data, will further strengthen the open collaborative culture in Sweden.
Planned approach and implementation
The project has three main parts: 1.A study of how sharing open data can be organized. We use Open Street Map as an example of how data is collected, organized, and used in various applications. We compare how open source communities work, such as Gerrit and Jenkins. 2.A survey of Swedish companies and public sector needs for data for different machine learning applications, if open data collaboration can address these needs. 3.A validation of the conclusions in the form of an open workshop