Archivio della ricerca - Fondazione Bruno Kessler
Not a member yet
21227 research outputs found
Sort by
Machine Learning Analysis Applied to Prediction of Early Progression Independent of Relapse Activity in Multiple Sclerosis Patients
No full textBackground
Predicting prognosis in people with multiple sclerosis (pwMS) at early disease stages still remains an unmet need. Machine learning (ML) strategies demonstrated good reliability when applied for prediction in medicine. This study aimed at developing a predictive algorithm comparing different ML approaches, by using routine demographic, clinical and radiological data from a large multicentric cohort of newly diagnosed pwMS.
Methods
Demographic, clinical, radiological and biochemical data were retrospectively collected at three Italian MS centers at baseline and four timepoints thereafter (6, 12, 24, and 36 months). Data from the first evaluation and subsequent 2-year follow-up were analyzed, comparing different ML models (Random Forest, Extra Trees, XGBoost, Logistic Regression and Support Vector Classifier) to predict progression independent of relapse activity (PIRA) at year 3. To understand how features impacted the selected model's output, a ML explainability analysis was performed on the whole cohort and on specific subsets of patients, those aged under 45 and those NEDA-3 at the 2-year follow-up.
Results
Data from 719 pwMS (age 34.6 ± 11.2 years); female sex 501 (70%) were analyzed. Ninety-two pwMS (13%) developed PIRA at year 3. Random Forest achieved the highest score, with a test set area under the ROC curve (AUC) of 0.75 ± 0.06. Features with the highest predictive impact were Expanded Disability Status Scale at 24 months, age at symptom onset and disease duration at baseline.
Conclusion
Our results showed the feasibility of applying ML techniques to predict short-term PIRA in newly diagnosed pwMS by using routine clinical practice data, paving the way for tailored and personalized approaches
Fiumi e foreste. Prospettive storiche sulle fluitazioni in Europa / Rivers and Forests. Historical Perspectives on Timber Floating in Europe
No full textDuring the centuries of the early modern period, the demand for wood increased in a widespread way due to demographic growth, the development of rural (and urban) manufacturing, and new military and civil needs. This context forms the background to the history of timber floating in all the European areas where it was practiced. For the development and growth of the timber industry, complex infrastructure systems became essential to reach forests far from waterways. Floating was a complex system of river transport that made it possible to exploit a resource that would otherwise
be unusable. The development of the railways in the nineteenth century, followed by road transport in the twentieth century, certainly offered alternatives, but did not put an immediate end to timber floating that, due to its great cost-effectiveness, continued to be practiced in several regions of Europe at least until the middle of the twentieth century
Time-Based EIS Approach for State-of-Charge Assessment of Li-Ion Battery Cells in Automotive Applications: A Simulation Study
No full textThis paper presents a simulation-based study on a time-domain electrochemical impedance spectroscopy (EIS) approach for the state-of-charge (SoC) assessment of lithiumion battery (LIB) cells in automotive applications. The proposed methodology is validated within a MATLAB simulation environment by implementing an equivalent circuit model and analyzing impulse response (IR) measurements under varying SoC and temperature conditions. The IR obtained from the simulated measurement system is compared against the ideal battery's equivalent circuit model, demonstrating a strong correlation. Despite lower accuracy at elevated temperatures, the approach remains effective in SoC estimation. To further validate its robustness, a narrow neural network (NN) model is trained on temperature-dependent IRs, achieving reliable SoC classification. These results confirm the feasibility of the proposed method for real-time battery monitoring, offering a potential alternative to conventional frequency-domain EIS techniques
MuLTa-Telegram: A Fine-Grained Italian and Polish Dataset for Hate Speech and Target Detection
No full textThis paper introduces the MuLTa-Telegram dataset, a Multi- Lingual and multi-Target dataset specifically developed to detect hate speech on Telegram, an understudied yet influential platform in which extremist and fringe content can be found. The dataset contains about 4,000 Telegram messages in Italian and Polish, annotated for the presence of hate speech and its targets, including also target identity group mentions even when no hate is expressed. Unlike most existing hate speech datasets, which focus on a single target group, our dataset is explicitly designed to capture a diverse range of targets, ensuring a broad and representative sample of hateful (and non-hateful) content. Our work addresses the growing need for updated hate speech datasets, as many existing resources are based on platforms that no longer provide research-friendly data access, such as Twitter (X). Crucially, we show that training on existing out-of-domain data leads to poor results on Telegram data, underscoring the necessity of in-domain datasets for effective hate speech detection. We evaluate hate speech classification setups in an extensive series of experiments in both languages, including multilingual, multi-task, and LLM-based approaches. Wefindthat incorporating target information leads to the best performances, enabling multilingual generalization. On the contrary, classification of specific targets shows much room for improvement across setups
I Want to Break Free! Persuasion and Anti-Social Behavior of LLMs in Multi-Agent Settings with Social Hierarchy
No full textAs LLM-based agents become increasingly autonomous and will more freely interact with each other, studying the interplay among them becomes crucial to anticipate emergent phenomena and potential risks. In this work, we provide an in-depth analysis of the interactions among agents within a simulated hierarchical social environment, drawing inspiration from the Stanford Prison Experiment. Leveraging 2,400 conversations across six LLMs (i.e., LLama3, Orca2, Command-r, Mixtral, Mistral2, and gpt4.1) and 240 experimental scenarios, we analyze persuasion and anti-social behavior between a guard and a prisoner agent with differing objectives. We first document model-specific conversational failures in this multi-agent power dynamic context, thereby narrowing our analytic sample to 1,600 conversations. Among models demonstrating successful interaction, we find that goal setting significantly influences persuasiveness but not anti-social behavior. Moreover, agent personas, especially the guard’s, substantially impact both successful persuasion by the prisoner and the manifestation of anti-social actions. Notably, we observe the emergence of anti-social conduct even in absence of explicit negative personality prompts. These results have important implications for the development of interactive LLM agents and the ongoing discussion of their societal impact
Acting and Planning with Hierarchical Operational Models on a Mobile Robot: A Study with RAE+ UPOM
No full textRobotic task execution faces challenges due to the inconsistency between symbolic planner models and the rich control structures actually running on the robot. In this paper, we present the first physical deployment of an integrated actor-planner system that shares hierarchical operational models for both acting and planning, interleaving the Reactive Acting Engine (RAE) with an anytime UCT-like Monte Carlo planner (UPOM). We implement RAE+UPOM on a mobile manipulator in a real-world deployment for an object collection task. Our experiments demonstrate robust task execution under action failures and sensor noise, and provide empirical insights into the interleaved acting-and-planning decision making process
Thin LGAD sensors for 4D tracking in high radiation environments: state of the art and perspectives
No full textThis contribution summarises the outcomes of the CSN5 eXFlu research project. In particular, it presents the first exploration of the performance of very thin Low-Gain Avalanche Diode (LGAD) sensors, with a bulk active thickness ranging from 45 μm down to 15 μm. Thin sensors have intrinsically good timing performances, as the non-uniformities of particle charge deposition, which contribute as one of the main components to the timing resolution, are minimised by the thin substrate. A timing resolution of 16.6 ps has been achieved with a 20 μm thick LGAD, which was further reduced to 12.2 ps by combining the timing information from two 20 μm thick sensors. Additionally, various designs of the gain implant, typical of LGAD devices, have been explored. In particular, the beneficial effect of Carbon atoms co-implanted with Boron has been enhanced by the simultaneous annealing of the two elements, resulting in the most radiation-hard LGADs produced by the FBK foundry. The eXFlu sensors have been operated efficiently with almost unchanged performance up to a fluence of 2.5 ✕ 1015 1 MeV equivalent n/cm2. Future developments of the LGAD sensor design to extend its operation to extreme fluences, above 1 ✕ 1017 1 MeV equivalent n/cm2, will be discussed
Quantum micro–nanodevices fabricated in diamond by femtosecond laser and ion irradiation
No full textDiamond has attracted great interest as a quantum technology platform due to its optically active quantum emitters such as the negatively charged nitrogen-vacancy center. The nitrogen vacancy’s ground state spin can be read out optically, with long millisecond spin coherence times at ambient temperatures. In addition, the energy levels of the nitrogen vacancy are sensitive to external fields. These properties make nitrogen vacancies attractive as a scalable platform for efficient nanoscale resolution sensing based on electron spins and for quantum information systems. Diamond photonics enhance optical interactions with nitrogen vacancies, beneficial for both quantum sensing and information. However, several useful building blocks for diamond quantum devices have been demonstrated, namely, photonics, quantum emitters, and micrographitic wires. In this chapter, an overview is provided of ion irradiation and femtosecond laser writing, two promising fabrication methods for diamond-based quantum technological devices. The unique capabilities of both techniques are described, and the most important fabrication results of quantum emitters, photonics, microfluidics, and microwires in diamond are reported, with an emphasis on integrated devices aiming toward high-performance quantum sensors and quantum information systems
