Protecting Data in Digital Agriculture

The Challenge:

Major agricultural producers are collecting increasing volumes of data, and face pressure in appropriately controlling and sharing these datasets with service providers and other industry bodies. Data and digital solutions can no longer be generated or developed on the scale required by individual companies.

Furthermore, the highly distributed and decentralized nature of IoT systems in agriculture makes proving data provenance and integrity problematic and threatens our sector’s ability to address the challenges of the 21st century.

‍

The Solution:

This project aims to overcome these broad global challenges of data privacy and security in three key areas of the data value chain:

• Sensing – Establish a framework for the inception of data that protects personally-identifiable information and complies with global anti-trust laws.
• Data Modelling (Federated Learning) - Evaluate best methods of federated learning in agriculture to enable data service providers to generate high-quality Machine Learning models from sensitive data aggregated from multiple data owners with low overheads in a confidential and privacy-preserving manner using Federated Learning.
• Exchange in data marketplaces - Explore mechanisms to create trusted digital cleanrooms for data marketplaces. This cleanroom service will mean that data owners can maintain ownership when uploading to a digital marketplace and retain full control and transparency over where and how the data is used.

Research activities will generate new insights and technologies that will be published broadly and adopted into Bosch’s global portfolio of data solutions, including other publicly funded projects like CRYPTECS or AgriGaia and open source initiatives that promote the applicability and adoption of privacy-preserving computing technologies, such as Carbyne Stack.

This project will continue to accelerate Bosch’s position as a world-leading data management specialist across agriculture and related sectors.

‍

Outcomes:

• Greater confidence for Australian farmers and other data providers in their ability to track and control their data
• Rapidly accelerated and streamlined applications of machine learning for data service providers across their broad base of customers
• Improved cohesion and performance of the digital agriculture ecosystem both in Australia and globally via trusted security and privacy-preserving solutions

‍

For more information, please contact hello@foodagility.com