Tackling risk prevention in food supply lines: Agroknow speaks on its involvement in the STELAR project
How do you feel about an innovative and easy-to-use platform that interlinks Big Data and sorts it into AI-ready data for usage in food safety per user requirements? Agroknow recognizes the importance of such a solution. This team of experts makes up a data and analytics company that uses AI to predict risks and prevent food recalls. As part of the EU project STELAR, they aim to redesign agrifood data analysis. How?
They are spearheading one of STELAR’s three pilots – Pilot A, for risk prevention in food supply lines. Seeing how agrifood data is currently scattered all over the place with diverse information such as regulation data, farm management data, price data, and weather data being stored in different formats etc., AgroKnow will aid the project’s technical team in creating aimed workflows by supplying the necessary data. That way, the project’s aforementioned platform – the Knowledge Lake Management System (KLMS) platform, is more likely to prove to be an innovative way of handling Big Data in risk prevention.
But we will give the floor to STELAR partners from Agroknow. Giannis Stoitsis, Manos Karvounis, and George Marinos will describe the disadvantages of traditional risk management, their proposed solution, the role farmers will play in Pilot A, as well as the current disruptions in the agrifood chain.
What food safety and food fraud-related issues cause the biggest disruptions in the agrifood chain?
On one hand we have the increasing complexity of the global food supply chain. On the other hand, we have global threats such as climate change and urbanization. All of this can lead to the emergence of food safety hazards that increasingly challenge the food supply chain.
Extreme events attributed to climate change are becoming more frequent, severe, and unpredictable. Such events impact food security by adversely affecting agricultural production and yield, disrupting supply chains. They also affect food safety. Elevated temperatures, alternation of severe drought periods and heavy rains, soil quality degradation, rising sea levels, and ocean acidification, among others, have implications for various biological and chemical contaminants in food. All these factors alter the contaminants’ virulence, occurrence, and distribution. That increases our risk of exposure to foodborne hazards. In addition, the rapid globalization of the food supply chains amplifies foodborne hazards, providing opportunities for local foodborne incidents to become international outbreaks, impacting food safety.
What were the biggest disadvantages of traditional risk management, and how can we change them?
The food system is rapidly changing, becoming increasingly digitized. Heterogeneous data is generated by all stakeholders (farmers, food companies, national authorities, consumers), including farm management data, weather data, environmental data, recalls data, border rejections data, laboratory testing data, regulation data, production data, trade data, price data, social data, product reviews data, inspection data, country risk data, audit data and others.
These are stored in local databases in different formats and languages, resulting in disconnected data silos. To perform a complete and timely risk assessment, understand emerging risks, and protect the consumers, domain experts need to be able to find, combine, process and extract meaning from as much data as possible across the entire supply chain. In this context, our focus is on how the STELAR KLMS can be used to enable the generation of integrated, AI-ready datasets for food risk prediction.
What innovative data management technologies can be harnessed to create unified, AI-ready datasets for predicting food risks?
STELAR will provide a KLMS for finding and interconnecting many data types from multiple diverse sources, eventually making them actionable. It will combine Big Data, Linked Data and Machine Learning technologies (proven effective in collecting and processing large and heterogeneous data) with the experience and knowledge of food safety experts. These professionals are indispensable for modeling, validating, classifying, and mining the data.
The proposed solution will unify and simplify access to the most critical data for risk assessment and prediction.
The KLMS is expected to produce rich and harmonized data. “Rich” means that all the needed information is available, e.g. type of hazard, type of product, traceability, links, and relations to other data sources, which can be predictors for the AI model. “Harmonized” means that all the required different datasets follow a common format and terminology so they can be correlated. That will allow food companies to manage and view all types of data of relevance and importance for risk prevention.
How do you plan on collaborating with partners on the project in order to contribute to building a knowledge lake for risk prevention in food supply lines?
AGROKNOW will support the STELAR technical team to create aimed workflows by providing all the necessary information and data. For example, the KLMS preferably needs 2+ years of incident frequency data of at least 100+ ingredients/hazards. Generally, the more data is provided to the KLMS, the more likely the innovative model is to obtain superior performance over the current methods.
Furthermore, to allow execution of the tasks, AGROKNOW will ensure that the end user:
- – Provides the description of each data source to be registered in the Data Catalog. This description will follow the Catalog’s API specification, including a few basic metadata and a set of optional metadata. The latter should include all the attributes needed to enable source prioritization later on.
- – Specifies their desired filtering and ranking criteria, including their relative importance.
- – Provides an appropriate crawler for each source to be crawled. The user should also indicate the desired schedule for crawling.
What role will farmers, food companies, and national authorities play in this pilot?
The farmers, food companies, and national authorities play the role of the end users whose needs will be addressed by integrating worldwide food safety related data sources.
STELAR KLMS will allow them to efficiently and timely discover the correct data for their needs.
What are the planned social and economic impacts, and how will Agroknow and the pilot reflect on the environment?
STELAR KLMS aims to contribute up to a 50% reduction of the time spent on important business decisions for risk prevention.
STELAR’s pilots will deal with data crucial across the entire food value chain. From optimizing food production and reducing its environmental footprint through precision farming to ensuring food safety for consumers through traceability, fraud detection, and risk analysis.
We will continue to describe STELAR’s pilots which will showcase how the KLMS assists in overcoming data management obstacles in the agrifood data analysis space. So don’t forget to check our Newsroom for more to come!