1,721,930 research outputs found
Comparing Generative AI Literature Reviews Versus Human-Led Systematic Literature Reviews: A Case Study on Big Data Research
No full textGenerative Artificial Intelligence (GenAI) and Large Language Models (LLMs) are transforming research methodologies, including Systematic Literature Reviews (SLRs). While traditional, human-led SLRs are labor-intensive, AI-driven approaches promise efficiency and scalability. However, the reliability and accuracy of AI-generated literature reviews remain uncertain. This study investigates the performance of GPT-4-powered Consensus in conducting an SLR on Big Data research, comparing its results with a manually conducted SLR. To evaluate Consensus, we analyzed its ability to detect relevant studies, extract key insights, and synthesize findings. Our human-led SLR identified 32 primary studies (PSs) and 207 related works, whereas Consensus detected 22 PSs, with 16 overlapping with the manual selection and 5 false positives. The AI-selected studies had an average citation count of 202 per study, significantly higher than the 64.4 citations per study in the manual SLR, indicating a possible bias toward highly cited papers. However, none of the 32 PSs selected manually were included in the AI-generated results, highlighting recall and selection accuracy limitations. Key findings reveal that Consensus accelerates literature retrieval but suffers from hallucinations, reference inaccuracies, and limited critical analysis. Specifically, it failed to capture nuanced research challenges and missed important application domains. Precision, recall, and F1 scores of the AI-selected studies were 76.2%, 38.1%, and 50.6%, respectively, demonstrating that while AI retrieves relevant papers with high precision, it lacks comprehensiveness. To mitigate these limitations, we propose a hybrid AI-human SLR framework, where AI enhances search efficiency while human reviewers ensure rigor and validity. While AI can support literature reviews, human oversight remains essential for ensuring accuracy and depth. Future research should assess AI-assisted SLRs across multiple disciplines to validate generalizability and explore domain-specific LLMs for improved performance
Editorial for the Special Issue on “Software Engineering and Data Science”, Volume II
Full text linkThe Special Issue “Software Engineering and Data Science, Volume II” is the natural continuation of its greatly successful predecessor, Volume I [...
Editorial for the Special Issue on “Software Engineering and Data Science”
No full textIn the last few years, data-driven software solutions have attracted a lot of attention in research and development at academic, industry, business, and government levels to exploit the hidden knowledge and big data that can be offered to cities and citizens in the future [...
Studying the Quality of Source Code Generated by Different AI Generative Engines: An Empirical Evaluation
No full textThe advent of Generative Artificial Intelligence is opening essential questions about whether and when AI will replace human abilities in accomplishing everyday tasks. This issue is particularly true in the domain of software development, where generative AI seems to have strong skills in solving coding problems and generating software source code. In this paper, an empirical evaluation of AI-generated source code is performed: three complex coding problems (selected from the exams for the Java Programming course at the University of Insubria) are prompted to three different Large Language Model (LLM) Engines, and the generated code is evaluated in its correctness and quality by means of human-implemented test suites and quality metrics. The experimentation shows that the three evaluated LLM engines are able to solve the three exams but with the constant supervision of software experts in performing these tasks. Currently, LLM engines need human-expert support to produce running code that is of good quality
Low-Cost, high sensitivity, signal processing-enhanced fiber Bragg grating sensing system for condition-based maintenance application
No full textThis paper presents a low-cost, high sensitivity fiber Bragg grating sensing system specifically designed to measure the amplitude of weak signals at audio frequencies. Low cost is obtained using for the optical part an intensity based interrogation scheme, with time and wavelength divi- sion multiplexing to create a sensor network that minimizes the expense per sensing point. Then, performances are dramatically improved using sequential signal processing techniques, which clear noise inside and outside the sensor, and provide an exceptionally performing spectral estimation. This way, the sensor gains 66 dB of signal-to-noise performance, and is able to measure vibra- tions as low as few picostrains, with a frequency resolution of 0.05 Hz. Example of application to multipoint condition-based maintenance application in industrial machineries is given. In this case, the proposed optical sensor is used to detect very weak signals buried into noise coming from ball bearings and rotating elements that are about to break, and can raise a warning alert well before critical damages occu
Understanding Artificial Intelligence in Chess: The RubiChess Case Study
No full textSoftware chess engines and advanced solutions based on Artificial Intelligence (AI) are changing the historical game of chess. Several AI-based chess engines are now available as closed or open software solutions with unique characteristics and game abilities. However, these AI-based engines are very complex to analyze from a functional point-of-view, and a deep study of their source code is not often enough to understand the advantages and disadvantages of their game approach. In this paper, we adopted a two-fold approach to study the behavior of a well-known AI-based chess engine called RubiChess. From one side, we studied the RubiChess architecture to understand its main game strategies and adopted algorithms. On the other side, we simulated a set of matches played against other AI chess engines, and we evaluated these matches (in different conditions) by using statistical tests and data visualization to determine the properties that made RubiChess unique. For example, the simulations highlight that RubiChess performs better as a white player and during "slow" games. This gives engine developers and players important insights into how RubiChess plays
An Artificial Intelligence-based tool to predict “unhealthy” wine and olive oil
No full textThe application of Artificial Intelligence (AI) in the agri-food industry has witnessed significant advancements, particularly in healthy food, quality assessment, and geographical origin determination of agri-products such as wine and olive oils, whose international market is constantly growing. This paper explores the applicability of Machine Learning (ML) models for predicting “unhealthy” wines and olive oils. As for the healthy factor, two main characteristics of wines and olive oils were evaluated: their quality and their geographical origin. In the study, 3 ML algorithms were compared (Random Forest, Linear Discriminant Analysis, and K-Nearest Neighbors) to predict just by observing their chemical characteristics: i) the quality of red and white wine and ii) the geographical origins of wine and olive oils. Real datasets were used for these case studies. The Synthetic Minority Over-sampling Technique (SMOTE) was used to manage imbalanced data, and well-adopted AI metrics were collected to evaluate the accuracy of the predictions made by the 3 ML algorithms. 3 datasets were analyzed: dataset #01 contains ∼7000 data entries related to 2 types of wine (red and white Portuguese wines) and 11 wine's chemical characteristics; dataset #02 contains 178 data entries related to 3 different Italian vineyards and 13 wine's chemical characteristics; dataset #03 contains 572 data entries related to 9 different Italian geographical regions for olive oil and 7 fatty acid characteristics. A total of ∼7700 data points, 5 wines and 9 olive oil types, and 31 chemical characteristics were analyzed. 12 quality metrics were collected per each dataset analysis. The designed ML models achieved good accuracy (>0.65) in predicting the wine quality, a very high accuracy (>0.99) in predicting its vineyard, and a very good accuracy (>0.97) in predicting the origin of olive oil. Moreover, an AI-based tool has been developed that stakeholders can use to predict the quality and geographical origin of wine and olive oil to avoid possible fraudulent and unhealthy product labels. The obtained prediction results showcase AI and ML's potential to help the agri-food industry, providing valuable insights and innovations for producers, distributors, and consumers alike. Integrating AI and ML in the agri-food industry may enhance product quality, consumer trust, and market transparency
- …
