Catalogue of Services

AI model evaluation and verification

Research Institutes of Sweden

Location

Remote

Sweden

Arable farming

Food processing

Greenhouse

Horticulture

Livestock farming

Tree Crops

Viticulture

Evaluating or verifying AI solutions is technically challenging and imposes specific requirements on the application, which can vary from setting internal benchmarks to establishing safety conditions. Evaluation and verification are performed using a broad set of tools ranging from straightforward performance evaluation metrics to formal methods for the verification of safety requirements. Formal methods can also be used to evaluate requirement fulfilment and systematically discover scenarios showing that AI systems need further improvements. Alternatively, customers can receive technical assistance and advice on improving AI systems that do not meet requirements.

AI model training

Politecnico di Milano

Location

Italy

Remote

Arable farming

Food processing

Greenhouse

Horticulture

Livestock farming

Tree Crops

Viticulture

This service concerns training AI models on behalf of the customer for a specific task and optimisation objective, e.g., improving accuracy on crop classification from image data. The target model is the solution provided by the customer that needs to be enhanced with respect to a set of pre-determined features to reach the desired performance level. However, if required, the training can also be applied to additional state-of-the-art models available in the market for benchmarking purposes. If not defined by the customer, some features of the training process can be identified via service S00179 (desk assessment activities for digital systems and/or data): for instance, model features to improve, reference model baselines to include in the performance comparison, as well as benchmark datasets. The data used for training the model can be either provided by the customer or annotated ad hoc as a preparatory activity to model training (via service S00290 - Data Labelling); another possibility is that data are retrieved among reference benchmark datasets that are openly available. We will also agree with customers on the level of hardware acceleration required, based on the considered AI models: e.g., GPU acceleration via connection to a remote server vs. on-device training.

AI model training for computer vision tasks in TEF infrastructure

GRADIANT

Location

Remote

Spain

Arable farming

Food processing

Greenhouse

Horticulture

Livestock farming

Tree Crops

Viticulture

The service allows training an AI model to be used for computer vision tasks, like object detection, segmentation, or classification tasks. The training is run on the TEF infrastructure, thus providing a powerful platform for developing cutting-edge artificial intelligence solutions in the agrifood sector. By leveraging high-performance computing resources, CPUs, and GPUs within the TEF infrastructure, this service enables innovators to train AI models using datasets based on images or videos. Users can either provide their own datasets or utilise existing data within the TEF infrastructure (see Related Services). The model training supports various architectures for the computer vision tasks, allowing the selection of the most appropriate one depending on the use case.

Annotated datasets for the validation of AI algorithms on crops, weeds, diseases, and pests

Wageningen Research - WR

Location

Netherlands

Remote

Arable farming

Horticulture

Tree Crops

In order to train, test, and validate AI applications, annotated datasets are needed. Wageningen Research - Field Crops can provide FAIR datasets consisting of high-quality raw images, annotations, and metadata on crops, weeds, diseases, and pests. At the moment a dataset on volunteer potato in sugar beet and onion is available. Let us know in which other datasets you are interested.

Assessment and evaluation of applied natural language processing

Research Institutes of Sweden

Location

Sweden

Remote

Arable farming

Food processing

Greenhouse

Horticulture

Livestock farming

Tree Crops

Viticulture

This service explores the potential of implementing applied Natural Language Processing (NLP) systems, rigorously evaluating them to ensure high quality and performance while identifying and addressing any limitations. It includes developing improvements to resolve issues uncovered during evaluation and ensures the NLP systems remain up-to-date.

By integrating NLP technology, businesses can enhance customer interactions through automated yet personalized responses. NLP also enables the analysis of large volumes of text data to uncover trends, identify opportunities, and provide valuable insights.

In the context of agricultural AI and robotics, NLP can significantly improve decision-making, operational efficiency, and communication.

Assessment of AI algorithm performance

Fondazione Bruno Kessler

Location

Remote

Arable farming

Greenhouse

Horticulture

Livestock farming

Tree Crops

Support clients in defining metrics and testing their AI solutions in real life, in the field, or virtually. Within normal and "stressed" operating conditions, including AI-generated test cases, we assess the readiness of agriculture-related AI algorithms for classification (weeds, chromatic analysis of crops), surround understanding and perception capabilities (for safety, for navigation), yield measurement (quantity of produce, quality), and robotics (anomaly detection, etc.). Besides accuracy metrics, we have experience evaluating memory and computing power footprints, and by doing so, understanding bottlenecks and improving the energy efficiency of proposed solutions.

Assessment of AI performance and software optimisation on edge computing devices

Fondazione Bruno Kessler

Location

At user's premises

Italy

Remote

Arable farming

Food processing

Greenhouse

Horticulture

Livestock farming

Tree Crops

Viticulture

The tests in this service are aimed at products that run complex apps (based on a multitude of microservices, including heavy AI algorithms) in the fields, with limited to no connectivity to the cloud. In these cases, we propose a benchmarking designed to understand what the bottlenecks are in the implemented solution architecture to help TEF customers improve the efficiency and reliability of such products/solutions. The service may also encompass virtual facility elements (enriched datasets, virtual machines hosting containerised modules) blended with physical sensors and infrastructure in our test field or directly installed at the client's premises if need be for the duration of the experimentation.

Assessment of dataset quality

Laboratoire National de Métrologie et d’Essais - LNE

Location

Remote

Arable farming

Food processing

Greenhouse

Horticulture

Livestock farming

Tree Crops

Viticulture

With the development of AI models in most sectors of society, the importance of the data used to train and test these models has been an important concern for the AI supplier. Understanding the quality of the datasets used, especially in sectors such as the agricultural and agrifood domains where data can be rare, is thus a prerequisite to ensure the correct development of AI models and build trust regarding these approaches for society. This service aims to answer this challenge by assessing the quality of datasets provided by the client. The features of the datasets that will be analyzed are based on agreed definition by the AI community: completeness, balance, diversity, accuracy... The dataset's quality assessment will begin by establishing the list of the system's influencing factors necessary for the constitution of a database. These factors will be identified both by using LNE expertise and outside sources, such as agrifood-relevant experts. Using these factors, LNE will analyse the distribution (or weight) of the various influencing factors in the databases and compare it to that under normal conditions of use in order to assess the representativity and balance of the dataset.

Assessment of interoperability for AI-driven solutions

Wageningen Research - WR

Location

At user's premises

Netherlands

Remote

Arable farming

Greenhouse

Horticulture

Livestock farming

Tree Crops

Viticulture

Data interoperability in the agrifood sector hinders innovation and development due to the need for many custom solutions to share data. This service helps agricultural organisations improve how they handle and share data across the entire food value chain. By implementing standardised reference data models like rmAgro (https://rmagro.org), we help to optimise your data flows and make your information systems work better together. Our team at Wageningen Research provides expert guidance in data modelling and implements reference models that align with industry standards and tests the reference models against various use cases. This promotes more efficient data sharing between different systems and organisations, reducing data integration challenges and improving operational efficiency. The service is particularly valuable for organisations looking to modernise their data infrastructure or needing to share data more effectively with partners in the agri-food sector. This service provides an assessment of interoperability for AI-driven solutions within the agri-food sector. The service facilitates conformance testing and verification of whether the related IT systems comply with relevant standards, guidelines, data space regulations, and other interoperability requirements. By evaluating the IT systems against established reference data models and frameworks like rmAgro (https://rmagro.org), the service ensures that it meets the necessary criteria for effective data sharing and integration. Wageningen Research leverages its expertise to evaluate the overall AI solution on its quality, performance, and how well it aligns with industry standards, offering insights and recommendations for improvement. This service supports organisations in ensuring their solutions are interoperable, compliant, and ready for seamless integration within the agri-food value chain.

Assessment of the feasibility of integrating payloads or equipment into UGVs

GRADIANT

Location

Remote

Spain

Arable farming

Greenhouse

Horticulture

Livestock farming

Tree Crops

Viticulture

This service will evaluate the feasibility of securely integrating third-party equipment or payloads related to the agrifood sector into unmanned ground vehicles (UGVs). The goal is to assess whether such integrations can be done safely and efficiently, ensuring that the added equipment does not interfere with the performance and operation of the UGV. This service will place special focus on key integration concepts such as minimum functional requirements, communication interfaces, power supply for equipment, physical integration of devices, and software compatibility. The outcome will include an evaluation of the feasibility of the integration, along with possible recommendations for the suitable configuration to ensure the proper functioning of the system.

Assessment of the feasibility of integrating payloads or equipment into UAVs

GRADIANT

Location

Remote

Spain

Arable farming

Horticulture

Livestock farming

Tree Crops

Viticulture

This service will evaluate the feasibility of securely integrating third-party equipment or payloads related to the agrifood sector into unmanned aerial vehicles (UAVs). The goal is to assess whether such integrations can be done safely and efficiently, ensuring that the added equipment does not interfere with the performance and operation of the UAV. This service will place special focus on key integration concepts such as minimum functional requirements, communication interfaces, power supply for equipment, physical integration of devices, and software compatibility. The outcome will include an evaluation of the feasibility of the integration, along with possible recommendations for the suitable configuration to ensure the proper functioning of the system.

Assistance for Identifying AI Technologies for Agricultural Robotics, Drones, and Precision Machinery

Danish Technological Institute

Location

Denmark

Remote

Arable farming

Food processing

Greenhouse

Horticulture

Our service is tailored to support AI technology providers, particularly in robotics and drone sectors, focusing on agricultural applications. We identify AI technologies that boost agricultural efficiency and productivity and offer integration support for embedding AI into agricultural robotics, drones, and precision machinery. Our guidance includes selecting effective AI tools and algorithms for specific agricultural challenges. Additionally, we offer comprehensive screening of AI technologies for agricultural relevance, which supports the company’s seamless integration of AI into existing equipment.