By the time the globe’s population has increased to the 9 billion it is predicted to in the year 2050, demands for water, food, energy and the threats surrounding climate change will present a host of supply challenges.
Image Credit: Arthimedes/Shutterstock.com
It is becoming increasingly apparent that isolated solutions to address challenges in just one industrial sector will not be adequate. This has led to the emergence of a concept called the food-water-energy nexus –interconnected networking of resources that will aid policymaking in the area of food production that will ultimately nourish the world.
Uptake of the technologies need is low
Precision farming or site-specific farming would be key to this, but farmers have not yet implemented the necessary technologies, the effect of which on-farm profitability is not yet clear. For example, the adoption of variable rate technology has been relatively low.
Applying this technology for the use of field fertilizer, for example, would require precise soil data across farms that are costly to acquire. Unless the profitability of site-specific management practices can be determined, uptake will remain low.
The adoption of such technologies would require proof of the associated potential yield profits and decreased run-off from the use of nitrogen and phosphorous for example would be needed.
Farmers dealing with livestock would need assurance that fewer hormones and antibiotics would be needed. This raises major concerns and questions about how digital agriculture, which incorporates precision farming, data science, and data ownership, could help to revolutionize agricultural throughput. Data science could be applied to link input factors to agricultural throughput in crop- and region-specific ways, whilst effectively harnessing the resources that would be available.
The possibility of using digital data in agriculture raises important questions about data privacy, protection and ownership, economic and technical logistics and sustainability.
Assessing the techniques, policies, and challenges
Now, in a comprehensive review of the literature covering the potential of using artificial intelligence in digital agriculture, Somali Chaterji (Department of Agricultural and Biological Engineering at Purdue University) and colleagues have addressed key questions surrounding the techniques, policies, and challenges faced.
This paper is the first one to bring together the important questions that will guide the end-to-end pipeline for the evolution of a new generation of digital agricultural solutions, driving the next revolution in agriculture and sustainability under one umbrella,”
Starting with the data generated from sensors, the team discusses data gathering and dissemination methods and data volume and quality before moving onto the various stages involved in the data’s lifecycle and the technologies needed for processing and interpreting analytics.
A section on how artificial intelligence could be democratized for farmers covers topics such as the availability and affordability of wireless and machine learning technologies and broadband modalities for data gathering.
For example, the digital agriculture technologies currently available are expensive due to a lack of innovation platforms and a lack of data that can be used to create machine learning models.
In this context, the authors propose “the building of open-source APIs [Application Programming Interfaces] and data repositories. These APIs, such as the one from Open Ag Data Alliance, afford researchers and startups a platform to prototype their innovation.”
A section on big agricultural data covers the frameworks and techniques that would be used to build machine learning algorithms and the constraints presented by stream processing being a preferred option to batch processing.
The authors also discuss the growing interest in carbon-neutral biofuels as alternative energy sources to fossil fuels and the optimization strategies that could be harnessed for advanced biofuel production.
Finally, the authors address important issues relating to policymaking for the regulation of big data and the economic factors involved in its adoption. The team says the most relevant policy questions concern the value and legal status of the farm and related business data.
“Academic research should focus on designing incentive-compatible mechanisms that encourage data sharing with public universities while protecting farmers' intellectual property,” suggests the team.
AI-Assisted Raman-Activated Cell Sorting for Microbial Analysis