AI for Sustainability: Activities of the CINI-AIIS Lab at University of Naples Federico II

Abstract

Sustainability is pivotal to global development, aligning closely with the United Nations' goals for a sustainable future. This paper introduces and discusses the perspectives and initiatives undertaken in these regards by the CINI AI-IS (the Italian National Consortium for Informatics, Artificial Intelligence and Intelligent Systems) Lab at the University of Naples Federico II. We will first introduce the DroughtScope project, currently on board the Kanyni Australian satellite to exploit hyperspectral data to detect early water stress in crops and optimize water resource management. We will then describe the PIVA project, addressing the challenge of missing data in complex systems, which occurs frequently in environmental domains, using Physics-Informed Variational Auto-Encoders to prevent model collapse. Additionally, the impact of Agriculture 4.0 on farmer health and workplace safety is discussed, examining the challenges and opportunities presented by advanced technologies. Finally, the paper considers the environmental and ethical implications of AI's carbon footprint, emphasizing the need for a balanced approach to technological advancement and environmental accountability.

Keywords

• eol>Synthetic data
• Carbon footprint
• Ethics
• Human-Centred AI

-

ronmental changes, and facilitating more informed decisions about natural resource management. AI’s capability to process and analyze vast amounts of environmental data enhances our ability to respond to climate change, manage natural disasters, and protect biodiversity. However, the application of AI in sustainability also raises important ethical and practical challenges, including the risk of increased energy consumption, potential biases in decision-making processes, and the implications for employment in traditional farming and environmental conservation roles. It is essential to address these challenges by developing AI solutions that are not only efective but also equitable and inclusive.

In this paper, we will thus introduce and discuss the perspectives and initiatives undertaken on responsible and reliable AI by the CINI AI-IS (the Italian National Consortium for Informatics, Artificial Intelligence and Intelligent Systems) Lab at the University of Naples Federico II, specifically focusing on the activities involving the members of the PICUS Lab1 as part of the AI-IS Node. To this aim, Section 2 describes the DroughtScope project, a finalist in the ESA’s OrbitalAI IMAGIN-e competition 2, using hyperspectral data to optimize water resource management through early detection of water stress in crops and the generation of alerts for risk areas. Section 3 discusses the use of AI for cows’ mastitis detection, an inflammatory condition of the udder causing critical issues for dairy milk and animal health. Section 4 describes the PIVA project, focusing on missing data imputation

1. Introduction

Advancements and integrative applications of artificial intelligence (AI) in agritech and environmental sustainability are becoming increasingly important as the global community seeks innovative solutions to pressing environmental challenges. The integration of AI technologies in agricultural and environmental contexts promises to enhance eficiency, reduce resource waste, and improve decision-making processes, aligning with several of the United Nations Sustainable Development Goals (SDGs), such as responsible consumption and production (Goal 12) and climate action (Goal 13) [ 1 ]. Artificial intelligence ofers a transformative potential for the agricultural sector by optimizing resource use and maximizing output, thereby addressing food security and economic sustainability. For instance, AI-driven systems can predict crop yields, monitor crop health through real-time data, and provide precise inputs regarding irrigation and fertilization, significantly reducing unnecessary resource expenditure and environmental impact.

In environmental management, AI technologies are critical in monitoring ecosystem health, predicting enviin complex systems by means of Physics-Informed Variational Auto-Encoders. Section 5 focuses on the challenges and opportunities for improving farmers’ well-being and productivity in the era of Agriculture 4.0. Finally, Section 6 analyses the AI environmental impact, particularly focusing on the carbon footprint of large-scale AI models. accuracy

2. DroughtScope project

0.8

The DroughtScope project, which appears among the finalists in the OrbitalAI IMAGIN-e competition organized by the European Space Agency (ESA), focuses on identifying early water stress conditions in crops to optimize water resource management. Utilizing hyperspectral data 3. Data Analysis over Mastitis from IMAGIN-e, DroughtScope estimates evapotranspira- Detection tion (ET) at a plot scale, enabling the creation of synthetic indicators for early detection of water stress. The project Mastitis is a critical issue for dairy milk and animal health. employs a multi-task deep-learning network to produce It is an inflammatory condition of the udder which efa real-time crop/no crop classification map and the Evap- fects economically reduces milk yield. Several methodorative Stress Index (ESI) product. The project uses open ologies, biological and new AI-based mastitis detection data from the ESA WorldCover dataset for ground truth use machine learning algorithms and artificial intelliin crop mapping. Additionally, ECOSTRESS data, with a gence to analyse data from multiple sources, such as milk spatial resolution of approximately 70 meters, are used production records, udder health parameters and senfor ESI and are rescaled to 45 meters to align with the sor readings, to identify patterns that indicate the onset input hyperspectral data. The DroughtScope’s architec- of mastitis in dairy cows. By continuously monitoring ture is designed to economize memory usage by sharing and analysing these data points, AI systems can provide a single encoder across multiple tasks and leveraging early and accurate detection of mastitis, enabling timely a feature-based Knowledge Distillation (KD) technique. intervention and optimisation of dairy herd health and This project is currently on board the Kanyni Australian satellite. Some preliminary results are reported in Table 1, while Figure 1 shows some predictions made by the model.

recall 0.7 precision 0.7 f1 0.7 model size productivity. In recent years, researchers have more and missing completely at random (MCAR), missing at ranmore using AI for detecting mastitis in the early stages of dom (MAR), and missing not at random (MNAR). MCAR the disease. In 2023, an automatic detection method for occurs when data missingness is entirely independent of dairy cow mastitis using the fusion of udder temperature observed or unobserved variables. MAR happens when and size features based on deep learning was proposed the probability of missingness is related to the observed [ 2 ]. The author used the YOLO7 model, centre network data but not the missing data itself. Finally, MNAR arises and Support Vector Machines in his research, showing when the missingness is related to the unobserved data. promising performance in the early detection of masti- Over the years, diferent approaches have been protis. In the same year, the Cina Agriculture University posed to perform missing data imputation, i.e. the reconshowed that thermal infrared technology combined with struction of missing pieces of information starting from a deep CLE-Unet model can significantly improved the available ones. As for several other domains, recently detection accuracy of mastitis [ 3 ]. More recently, deep Deep Learning (DL) solutions are becoming more frelearning combined with bufalo udder ultrasound was quent. However, a significant challenge with DL-based used for the first time to detect mastitis with the aim of systems, such as those based on Variational Auto-Encoder establishing an accurate, fast and inexpensive method (VA) [ 7, 8 ], is their susceptibility to model collapse, where to detect bufalo mastitis instead of routine laboratory results often converge to median values. As this would examination [ 4 ]. The model is suitable for mastitis detec- make the imputed data statistically not compliant with tion and can be used in a small dairy farm, but still needs the underlying physical phenomena, it is crucial to demore deep learning methods to detect the disease more velop systems robust against this issue. ifnely. Besides the reported example, Table 2 reports and The PIVA (Physics Informed Variational Auto-Encoder) details other approaches recently used for the task. project is a solution specifically designed to prevent model collapse. To this aim, PIVA incorporates a dual mechanism approach, utilizing both physical constraints 4. The Piva project and a masking technique. It is structured on a Variational Auto-Encoder (VAE) architecture where constraints are In the era of data-driven decision-making, the quality and integrated into the loss function to guide the network tocompleteness of data play a crucial role across diverse wards adhering to essential physical and statistical paramifelds, ranging from medical to industrial and environ- eters. These include entropy conservation, summation mental applications. In all these cases, missing data is a constraints across variable groups, control over covarifrequent problem that can arise for various reasons, in- ances, and adherence to the Wasserstein distance. The cluding sensor malfunctions or human errors. Based on technique of conserving entropy in both generated and the particular underlying reason, the phenomena of miss- observed variables has been demonstrated to be particuing data are categorized into three distinct mechanisms: larly efective in mitigating model collapse. Additionally,

PIVA adopts a novel data masking strategy inspired by

the method used in Bert. Unlike traditional Denoising Auto-Encoders that focus on reconstructing data from noise-altered inputs, PIVA’s strategy concentrates on accurately predicting masked data while ensuring compliance with the constraints imposed on these data points. This approach is pivotal for achieving dependable data imputation and enhancing the overall robustness of the model.

5. Farmer Health in the Era of

Agriculture 4.0: Challenges and Opportunities

cycle of nature, of contributing to food production and community well-being, can be shaken when machines begin to perform these roles, leading to a sense of loss of identity, which can have negative efects on farmers’ mental health. Also, the risk of social isolation due to decreased human interaction can have a significant impact on mental health, causing feelings of loneliness and depression. Although Agriculture 4.0 undoubtedly brings advantages in terms of production and sustainability, it is essential to recognize and address the potential negative efects on farmers’ health, particularly those of a psychological nature.

To ensure the prudent and eficient development of 4.0 solutions in agriculture, regulations and directives should be implemented to ensure a safe and healthy working environment and promote the physical and mental wellbeing of agricultural workers. At the same time, it is essential to develop psychological support programs, provide resources to address change and promote a culture of mental well-being within agricultural communities to preserve the health and well-being of agricultural workers. The transition to Agriculture 4.0 ofers undeniable advantages in terms of eficiency and sustainability, but it is essential to consider the ethical implications in terms of social justice, environmental sustainability, and farmer health. Only through a multidisciplinary, fair, and sustainable approach, it may be possible to fully realize the potential of this technological revolution in agriculture, ensuring that the benefits are shared fairly and responsibly by all members of society.

In the context of Industry 4.0, Agriculture 4.0 represents a crucial step in the evolution of precision agriculture. Automation, the use of drones and sensors, data collection, and artificial intelligence have enabled a more precise and eficient approach to agriculture, promising to increase production and reduce waste. Moreover, 4.0 solutions emerge in response to climate change, contributing to mitigating its negative efects on crop yield, management dificulties, and farmer well-being [ 9 ]. Undoubtedly, the advent of Agriculture 4.0 marks a significant breakthrough in the agricultural sector, traditionally characterized by a strong reliance on manual outdoor labour, presenting itself as a valuable ally for those who work the land. For example, thanks to automation and the use of advanced agricultural machinery, farmers can reduce their direct exposure to adverse weather conditions. If autonomous tractors and drones allow work in the fields 6. The Carbon Footprint of AI: even in the presence of heavy rains, extreme temperatures, or excessive heat, thus reducing the risk of weather- Ethics and Environmental related diseases, optimizing agricultural processes allows Accountability farmers to also plan their activities more intelligently, avoiding the hottest hours of the day or adverse weather The widespread introduction of artificial intelligence at conditions [ 10 ]. Similarly, in terms of workplace safety, virtually every societal level prompts deep reflection on a particularly delicate issue in the industrial context, au- the consequences of the massive use of these technolotomation and the use of advanced machinery reduces the gies. This raises fundamental ethical questions about risk of accidents [ 11 ], thereby contributing to preserving how we should regulate and manage these innovations the health and lives of farmers. to ensure a positive impact on society as a whole. A

However, while these advancements undoubtedly ofer highly topical issue is the environmental impact of AI, benefits in terms of production and sustainability, it is particularly the carbon footprint of learning models. The essential to recognize and address the potential negative increase in the size of artificial intelligence models, espeefects on farmers’ health, particularly those of a psycho- cially those based on deep neural networks (DNNs), conlogical nature. The gradual replacement of traditional sequently results in higher energy consumption during tasks of agricultural workers with machines and auto- the training process [ 12 ]. This phenomenon is driven by mated systems, while increasing production and reducing the need for larger models to achieve better performance physically demanding work, could easily create a sense but raises concerns about the environmental impact due of alienation for those who have dedicated themselves to increased energy consumption. The fundamental questo agriculture for generations. The sense of personal and tion revolves around striking a balance between AI’s precultural identity is often deeply connected to agricultural cision goals and the environmental impact resulting from work. The perception of being an essential part of the such research. Essentially, to what extent is it ethical to pursue AI research and development focusing solely on model accuracy if it entails increased energy costs and pollution? It’s a moral trade-of that requires careful balancing: on one hand, the accuracy of AI models is crucial for many applications; on the other hand, the rise in energy costs and pollution raises ethical concerns about the sustainability of this approach. Furthermore, complicating this equation is the awareness that much of AI’s energy costs come from the operational use of models [ 13 ], highlighting the environmental responsibility not only of researchers and developers but also of AI companies, energy providers, and various stakeholders involved. Identifying and fairly distributing responsibilities among the various actors involved can thus be challenging, as they may have conflicting interests and viewpoints. Developers may focus on innovation and model accuracy, while AI companies may be incentivized to maximize profits, ignoring environmental impacts. Additionally, energy providers may resist transitioning to more sustainable energy sources for economic reasons.

These considerations underscore the need for a thorough reflection on responsible resource management by society as a whole to mitigate the environmental impact of AI from the perspective of ethical and sustainable progress. The consequences of artificial intelligence do not only concern the technical field. Today, ethics therefore play a central role in addressing the various challenges posed by artificial intelligence, and balancing technological progress with environmental responsibility is a particularly delicate moral issue. What seems desirable and necessary is a coordinated efort by industry, academia, governments, and civil society to find balanced solutions that take into account both technological progress and sustainability needs [ 14 ]. However, this approach requires open and transparent dialogue among all stakeholders, as well as targeted policies and incentives that promote environmental and social responsibility in technological innovation. For example, international organizations and regulatory authorities can collaborate to develop sustainability standards for AI, including environmental criteria to be respected during the development, implementation, and use of AI models (an example is the International Telecommunication Union, ITU, which has established a working group on environmental issues related to AI, tasked with developing recommendations and guidelines to promote sustainable use of the technology) [ 15 ]. In an era where technological innovation is advancing at an unprecedented pace, it is essential to consider the long-term implications for the environment and society. The adoption of AI ofers enormous benefits in terms of eficiency, automation, and performance improvement, but it must be guided by ethical values and principles of social equity. This is because the decisions made today regarding the development and implementation of AI will have a significant impact on our planet and future generations. Furthermore, it is important to adopt a proactive approach in defining policies and regulations that guide the responsible development and use of AI, ensuring that the interests of society as a whole are adequately represented through a holistic and collaborative approach.

Acknowledgments

This work was partially supported by PNRR MUR Project PE0000013-FAIR.

References

[1] M. Bexell , K. Jönsson , Responsibility and the united nations' sustainable development goals, in: Forum for development studies , volume 44 , Taylor & Francis, 2017 , pp. 13 - 29 .

[2] M. C. Q. L. Y. W. X. Z. Y. .GangLiu , Fusion of uddar temprature and size features for the automatic detection of dairy cow mastitis using deep learning , Computers and Electronics in Agriculture 212 ( 2023 ). doi: 10 .1016/j.compag. 2023 . 108131 .

[3] A. K. M. S. S. Sachdeva , Detection of mastitis disease in cow with machine learning classifiers , Research Gate 7 ( 2023 ) 112 - 129 . doi: 10 .52865/DQFZ601.

[4] X. J. Y. L. Y. Z. Z. Y. W. Z. P. N. L. Yang , A new method to detect bufualo mastitis using uddar ultrasonography based on deep learning , MDPI 14 ( 2024 ) 707 . doi:https://www.mdpi. com/2076-2615/14/5/707#.

[5] Q. Z. Y. Y. G. L. Y. N. J. Li , Dairy cow mastitis detection by thermal infrared images based on cle− unet , MDPI 13 ( 2023 ) 2211 . doi:https: //www.mdpi.com/2076-2615/13/13/2211#.

[6] T. B. R. G. A. G. S.Befani, Exploiting machine learning methods with monthly routine milk recording data and climate information to predict subclinical mastitis in italian mediterranean buffaloes , NIH 106 ( 2023 ) 1942 - 1952 . doi: 10 .3168/ jds.2022- 22292 .

[7] Y. Sun , J. Li , Y. Xu , T. Zhang , X. Wang , Deep learning versus conventional methods for missing data imputation: A review and comparative study , Expert Systems with Applications ( 2023 ) 120201 .

[8] J. Yoon , J. Jordon , M. Schaar , Gain: Missing data imputation using generative adversarial nets , in: International conference on machine learning, PMLR , 2018 , pp. 5689 - 5698 .

[9] D. C. Rose , R. Wheeler , M. Winter , M. Lobley , C. -A. Chivers, Agriculture 4.0: Making it work for people, production, and the planet , Land use policy 100 ( 2021 ) 104933 .

[10] V. Goltyapin , I. Golubev , Global trends in the development of monitoring systems for mobile agricultural equipment , in: E3S Web of Conferences , volume 157 , EDP Sciences , 2020 , p. 01013 .

[11] N. Stacey , P. Ellwood , S. Bradbrook , J. Reynolds , H. Williams , D. Lye , Foresight on new and emerging occupational safety and health risks associated with digitalisation by 2025, Luxembourg: European Agency for Safety and Health at Work ( 2018 ).

[12] R. Schwartz , J. Dodge , N. A. Smith , O. Etzioni , Green ai , Communications of the ACM 63 ( 2020 ) 54 - 63 .

[13] D. Patterson , J. Gonzalez , Q. Le , C. Liang , L. -M. Munguia , D. Rothchild , D. So , M. Texier , J. Dean , Carbon emissions and large neural network training , arXiv preprint arXiv:2104.10350 ( 2021 ).

[14] G. Tamburrini, The ai carbon footprint and responsibilities of ai scientists , Philosophies 7 ( 2022 ) 4 .

[15] H.-T. Liao , C.-L. Pan , Y. Zhang, Smart digital platforms for carbon neutral management and services: Business models based on itu standards for green digital transformation , Frontiers in Ecology and Evolution 11 ( 2023 ) 1134381 .