Fabiano Pecorelli

Timely and accurate identification of crop diseases is a critical challenge in precision agriculture, as delayed or incorrect diagnosis often leads to excessive pesticide use, yield loss, and environmental degradation. Existing vision-based disease detection systems primarily focus on classification accuracy but lack interpretability and decision support, while large language models (LLMs) suffer from hallucinations when deployed without domain grounding. To address these limitations, this paper proposes a knowledge-grounded vision–language framework that connects YOLOv8-based disease detection with a Retrieval-Augmented Generation (RAG) pipeline and a Large Language Model (LLM). The framework transforms structured visual detection outputs into retrieval queries that guide knowledge-grounded reasoning, enabling fact-consistent explanations and disease-specific remediation recommendations while mitigating hallucinations. Experiments conducted on the BRACOL coffee leaf dataset demonstrate that the proposed system achieves strong detection performance using YOLOv8n, attaining an overall mAP@0.5 of 0.792 and mAP@0.5:0.95 of 0.523 on the original annotations. Compared to refined annotations, the original dataset yields consistent performance gains across all disease classes, highlighting the sensitivity of detection models to annotation strictness. Furthermore, a Non-Maximum Suppression (NMS) based ensemble of multiple YOLO variants improves robustness on challenging classes, increasing mAP@0.5 from 0.619 (best single model) to 0.634 on the reviewed dataset. By combining accurate visual detection with knowledge-grounded language generation, the proposed framework provides an interpretable, user-friendly decision support tool for farmers. This work demonstrates the potential of integrating object detection, RAG, and LLMs to support sustainable, data-driven disease management and reduce unnecessary pesticide usage in precision agriculture.

Software sustainability is a key multifaceted non-functional requirement that encompasses environmental, social, and economic concerns, yet its integration into the development of Machine Learning (ML)-enabled systems remains an open challenge. While previous research has explored high-level sustainability principles and policy recommendations, limited empirical evidence exists on how sustainability is practically managed in ML workflows. Existing studies predominantly focus on environmental sustainability, e.g., carbon footprint reduction, while missing the broader spectrum of sustainability dimensions and the challenges practitioners face in real-world settings . To address this gap, we conduct an empirical study to characterize sustainability in ML-enabled systems from a practitioner’s perspective. We investigate (1) how ML engineers perceive and describe sustainability, (2) the software engineering practices they adopt to support it, and (3) the key challenges hindering its adoption. We first perform a qualitative analysis based on interviews with eight experienced ML engineers, followed by a large-scale quantitative survey with 203 ML practitioners. Our key findings reveal a significant disconnection between sustainability awareness and its systematic implementation, highlighting the need for more structured guidelines, measurement frameworks, and regulatory support.

Neurosymbolic AI is an emerging paradigm that combines neural network learning capabilities with the structured reasoning capacity of symbolic systems. Although machine learning has achieved cutting-edge outcomes in diverse fields, including healthcare, agriculture, and environmental science, it has potential limitations. Machine learning and neural models excel at identifying intricate data patterns, yet they often lack transparency, depend on large labelled datasets, and face challenges with logical reasoning and tasks that require explainability. These challenges reduce their reliability in high-stakes applications such as healthcare. To address these limitations, we propose a hybrid framework that integrates symbolic knowledge expressed in First-Order Logic into neural learning via a Logic Tensor Network (LTN). In this framework, expert-defined medical rules are embedded as logical axioms with learnable thresholds. As a result, the model gains predictive power, interpretability, and explainability through reasoning over the logical rules. We have utilized this neurosymbolic method for predicting diabetes by employing the Pima Indians Diabetes Dataset. Our experimental setup evaluates the LTN-based model against several conventional methods, including Support Vector Machines (SVM), Logistic Regression (LR), K-Nearest Neighbors (K-NN), Random Forest Classifiers (RF), Naive Bayes (NB), and a Standalone Neural Network (NN). The findings demonstrate that the neurosymbolic framework not only surpasses traditional models in predictive accuracy but also offers improved explainability and robustness. Notably, the LTN-based neurosymbolic framework achieves an excellent balance between recall and precision, along with a higher AUC-ROC score. These results underscore its potential for trustworthy medical diagnostics. This work highlights how integrating symbolic reasoning with data-driven models can bridge the gap between explainability, interpretability, and performance, offering a promising direction for AI systems in domains where both accuracy and explainability are critical.

This paper presents a qualitative study on the barriers to the acceptance and development of quantum computing and the remedies for mitigating them. We interviewed eight quantum stakeholders from various organizations in the Netherlands and Germany with different levels of quantum expertise and diverse incentives to engage with technology development. We found eleven barriers to the acceptance and growth of quantum computing that reflect the limitations and uncertainty in the technology and the challenges to knowledge transfer and collaboration among academia, industry, policymakers, and society (i.e., the quantum stakeholders). We also found nine possible remedies to overcome these barriers by enabling effective collaboration and unobstructed information flow among stakeholders. Overall, we contribute to the existing literature on quantum computing adoption by identifying obstacles and remedies that consider both technological and stakeholder perspectives.

Test flakiness arises when a test case exhibits inconsistent behavior by alternating between passing and failing states when executed against the same code. Previous research showed the significance of the problem in practice, proposing empirical studies into the nature of flakiness and automated techniques for its detection. Machine learning models emerged as a promising approach for flaky test prediction. However, existing research has predominantly focused on within-project scenarios, where models are trained and tested using data from a single project. On the contrary, little is known about how flaky test prediction models may be adapted to software projects lacking sufficient historical data for effective prediction. In this paper, we address this gap by proposing a large-scale assessment of flaky test prediction in cross-project scenarios, i.e., in situations where predictive models are trained using data coming from external projects. Leveraging a dataset of 1,385 flaky tests from 29 open-source projects, we examine static test flakiness prediction models and evaluate feature- and instance-based filtering methods for cross-project predictions. Our study underscores the difficulties in utilizing cross-project flaky test data and underscores the significance of filtering methods in enhancing prediction accuracy. Notably, we find that the TrAdaBoost filtering method significantly reduces data heterogeneity, leading to an F-Measure of 70%.

Quantum computing is an emerging field in which theoretical principles are being transformed into practical applications, largely due to the efforts of the developer community. In order to ensure that quantum software engineering continues to advance, it is vital to understand the experiences, challenges, and aspirations of developers. This chapter is a continuation of our previous work, which provided a comprehensive survey exploring the adoption patterns and common challenges in quantum software engineering. In addition to the survey, we conducted in-depth, semi-structured interviews with practitioners in the field to gain a deeper and more detailed understanding of their perspectives. Through the interviews and survey findings, we have gained nuanced insights into the motivations, hurdles, and outlook of developers toward the rapidly evolving quantum computing landscape. We describe the research methodology in detail, including the tools and techniques used, in order to provide a comprehensive understanding of the research process. Furthermore, we present critical insights from both the survey and interviews, enriching the narrative with fresh perspectives obtained from the post-publication interviews. This chapter is a blend of academic investigation and real-world practitioner insights, aiming to provide a comprehensive understanding of the current state of quantum software engineering. By illuminating the path for future research and development in this dynamic field, we hope to guide the way toward continued progress and innovation.