Project Outline

Motivation

The agricultural sector has been increasingly digitalised by Artificial Intelligence (AI) and Internet-of-Things (IoT) for enhancing crop production modelling and irrigation management decisions to amplify yield and optimise agricultural resources. Leveraging AI (including machine learning, computer vision, and data analytics) and IoT (including real-time sensor and drones) can automate many data-driven intelligent agricultural processes. However, digitalisation of agriculture will also introduce unprecedented cybersecurity threats, jeopardising the $80 billion sector.  

The future of Australia’s agriculture will be heavily dependent on intelligent technologies like AI and IoT to a large extent for a variety of tasks, such as crop production, irrigation management and disease diagnosis. However, typical agricultural participants are inadequately prepared for and vulnerable to the imminent cyber attacks threatening AI applications at the emerging phase, as has been observed in other sectors. It is paramount to develop user-friendly and automated solutions to help average farmers mitigate the future cyber attacks that will cripple their livelihoods.

At Data61, our project “SCATES: Securing Critical Agriculture Technology and Emerging Solutions” (funded by Cyber Security CRC) aims to provide user-friendly cyber secure solutions for agriculture to average farmers. This PhD scholarship will leverage the resources and connections to the farming communities and government bodies in the SCATES project. Furthermore, the connection to ANU will provide the student with the necessary research knowledge and academic support in the relevant areas.

Methodology

The project will focus on the possible attack models and mitigation strategies for the following two scenarios:
Adversarial attacks (e.g., adversarial example, poisoning attack and backdoor attacks) on machine learning/AI models for crop production, irrigation management  decisions, and disease diagnosis.
Cyberattacks on wireless IoT devices, especially when the AI is incorporated. For example, such attacks could manipulate the data sensed by the IoT devices, which can allow attackers to hijack or influence IoT sensors for malicious purposes.

Deliverables

In this project, the student will deliver

• Tools to detect and mitigate attacks in ML models and IoT sensors. These tools are built based on the novel findings of the project and are implemented and made available in the desired format for the target community, for example the end users or research community.
• Reports and publications of analysis to digitalised agriculture empowered by secure AI and IoT. The outcomes will be published in top-tier multidisciplinary AI conferences or journals. The findings will also be distributed to the public and the relevant government/industry entities in non-technical terms.

Engagement Plan

We plan to actively engage with CSIRO and government stakeholders, including federal/NSW government, through various channels during the life of the PhD program. The goal is to continuously communicate our findings and receive feedback. Once some of the findings of the project are mature, we hope to explore the pathways to apply secure smart agriculture to the local farming communities.

The student will learn

The student will learn about the state-of-the-art tools and techniques in AI, cybersecurity, and digitalisation of agriculture, supported by a diverse strong panel of supervisors. They will have the opportunity to gain valuable insights into cybersecurity threats and solutions to digital agriculture and broad intelligent applications of agricultural data management. They also will have a chance to implement practical solutions for local farming communities. Last but not least, the student will improve their analytical, problem-solving, and management skills through the program.

To register an expression of interest, click here. You will need to outline why you have selected the research project and how your skills, experience and/or knowledge meet the project requirements.