arXiv subjects

Italo Epicoco

Publications and source records attributed to Italo Epicoco.

At least 19 recordsLinked to original sources

arXiv2026

Neptune: An AI model for Global Ocean Subseasonal Prediction

Subseasonal-to-seasonal (S2S) forecasting is societally critical, supporting decision-making in sectors ranging from water and agricultural management to disaster risk reduction, energy planning, and insurance. Achieving reliable predictions at these timescales requires representing the ocean and its dynamics, but traditional physics-based Ocean General Circulation Models (OGCMs), are computationally expensive and difficult to develop and improve because of the code complexity. In this work, we propose Neptune, an end-to-end data-driven framework for global ocean and sea-ice components emulation tailored for S2S timescales, up to 60 days. Neptune combines Convolutional Neural Networks (CNNs) and Spherical Fourier Neural Operators (SFNOs) to effectively capture local features and global cross-scale interactions, thereby obtaining a coherent representation of the ocean state. Forced by prescribed daily atmospheric fields, Neptune emulates ocean state variables, from temperature and salinity, to zonal and meridional currents, from sea surface height to sea ice thickness and concentration, with daily outputs at the ocean surface and through the water column. Specifically, we propose two variants of Neptune, Neptune-1 and Neptune-025, capable of emulating the ocean state at 1{\deg} and 0.25{\deg} resolution, respectively. Evaluated against a suite of metrics, including statistics (RMSE, CRPS and ACC), physical coherency (Ocean Heat Content, Eddy Kinetic Energy and Ice Brier Score) and climate indices (ENSO and Z20 metric, IOD), Neptune successfully reproduces the spatio-temporal evolution of the oceanic fields up to 60 days, and is stable over long timescales. Neptune provides compelling evidence that end-to-end data-driven ocean emulators can become a powerful component of next-generation S2S forecasting systems, emulating ocean state at high spatio-temporal resolution.

physics.ao-ph↗

arXiv2026

Fast and Accurate Anomaly Detection in Time Series

Anomaly detection is a critical and evolving field in Machine Learning, with applications targeting different domains such as cybersecurity, finance, healthcare, manufacturing and IoT (Internet of Things) systems. Traditionally, anomaly detection algorithms have been designed using both supervised and unsupervised learning paradigms. The fundamental challenge in real-world anomaly detection scenarios is related to the inherent class imbalance (anomalies are typically rare) and, for supervised methods, to the scarcity of labelled anomalous data. Indeed, labelling is both expensive and time-consuming. Conversely unsupervised methods do not require labelling, but may suffer from high false positive rates when deployed in safety-critical applications. In this work we introduce a novel unsupervised algorithm for anomaly detection in time series based on the Haar discrete wavelet and a suitably designed $t$-test. We establish the theoretical foundation of the proposed $t$-test and, through extensive experimentation across 343 datasets, demonstrate that our algorithm outperforms state-of-the-art unsupervised and self-supervised benchmarks.

cs.LG↗

arXiv2026

Operational evaluation of data-driven forest fire forecasting models

A growing body of literature has focused on predicting wildfire occurrence using machine learning methods, capitalizing on high-resolution data and fire predictors that canonical process-based frameworks largely ignore. Standard evaluation metrics for an ML classifier, while important, provide a potentially limited measure of the model's operational performance for the Fire Danger Index (FDI) forecast. Furthermore, model evaluation is frequently conducted without adequately accounting for false positive rates, despite their critical relevance in operational contexts. In this paper, we revisit the daily FDI model evaluation paradigm and propose a novel method for evaluating a forest fire forecasting model that is aligned with real-world decision-making. Furthermore, we systematically assess performance in accurately predicting fire activity and the false positives (false alarms). We further demonstrate that an ensemble of ML models improves both fire identification and reduces false positives.

cs.LG↗

arXiv2026

OpFML: Pipeline for ML-based Operational Inference

Machine learning models for climate and Earth science are becoming increasingly capable, yet model deployment into operational use remains a largely unaddressed challenge: general-purpose model-serving tools, such as MLflow and KServe, assume input data availability at the inference node, while data acquisition, failure handling, and preprocessing are trusted to a separate workflow. We present OpFML: Operational Forecasting with Machine Learning - a configurable pipeline integrating the four steps of operational inference into a single TOML-configured workflow: data consumption, contingency handling, preprocessing, and model inference. By consolidating these steps, OpFML removes the significant boilerplate code required for each new deployment. We demonstrate the pipeline on the operational forecasting of daily fire activity over southern Italy.

cs.LG↗

arXiv2025

MedFormer: a data-driven model for forecasting the Mediterranean Sea

Accurate ocean forecasting is essential for supporting a wide range of marine applications. Recent advances in artificial intelligence have highlighted the potential of data-driven models to outperform traditional numerical approaches, particularly in atmospheric weather forecasting. However, extending these methods to ocean systems remains challenging due to their inherently slower dynamics and complex boundary conditions. In this work, we present MedFormer, a fully data-driven deep learning model specifically designed for medium-range ocean forecasting in the Mediterranean Sea. MedFormer is based on a U-Net architecture augmented with 3D attention mechanisms and operates at a high horizontal resolution of 1/24°. The model is trained on 20 years of daily ocean reanalysis data and fine-tuned with high-resolution operational analyses. It generates 9-day forecasts using an autoregressive strategy. The model leverages both historical ocean states and atmospheric forcings, making it well-suited for operational use. We benchmark MedFormer against the state-of-the-art Mediterranean Forecasting System (MedFS), developed at Euro-Mediterranean Center on Climate Change (CMCC), using both analysis data and independent observations. The forecast skills, evaluated with the Root Mean Squared Difference and the Anomaly Correlation Coefficient, indicate that MedFormer consistently outperforms MedFS across key 3D ocean variables. These findings underscore the potential of data-driven approaches like MedFormer to complement, or even surpass, traditional numerical ocean forecasting systems in both accuracy and computational efficiency.

physics.ao-ph↗

arXiv2025

Application of association rule mining to assess forest species distribution in Italy considering abiotic and biotic factors

Biodiversity monitoring represents a pressing global priority, and assessing forest community composition plays a crucial role due to its influence on ecosystem functions. The spatial distribution of forest species becomes essential for understanding biodiversity dynamics, territorial planning, aiding nature conservation and enhancing ecosystem resilience amid global change. Association Rule Mining, commonly applied to other scientific contexts, is now innovatively adopted in the ecological field to explore the relationships among co-occurring plant species and extract hidden interpretable patterns, also with abiotic and biotic conditions. Multiple heterogeneous data sources were integrated through data preprocessing into a unique dataset, including georeferenced information about 151 plant species monitored within 6,784 plots across Italy and several bioclimatic indices, soil-related factors, and variables from earth observations. The Frequent Pattern Growth algorithm, used for association rule mining, provided interesting and encouraging findings, suggesting ecological rules among plant species and environmental conditions. Indeed, temperature seasonality between 650-700 and precipitation seasonality between 45-50 resulted very correlated with Picea abies (confidence = 90.9%, lift = 7.13). Patterns detected for Picea abies highlighted its ecological specificity, indicating a strong association with cold, highly seasonal environments, and particular plant communities. Some species appeared acting as community "hubs", frequently co-occurring with other species, suggesting ties to specific environmental or biotic conditions. These findings represent a valuable resource for future research, especially in regions with similar environmental settings and when prior ecological knowledge exists, also underlining the importance of publicly accessible, high-quality ecological data.

cs.CE↗

arXiv2025

Distributed P2P quantile tracking with relative value error

In this paper we present \textsc{DUDDSketch}, a distributed version of the \textsc{UDDSketch} algorithm for accurate tracking of quantiles. The algorithm is a fully decentralized, gossip-based distributed protocol working in the context of unstructured P2P networks. We discuss the algorithm's design and formally prove its correctness. We also show, through extensive experimental results, that the algorithm converges to the results provided by the sequential algorithm, which is a fundamental and highly desirable property.

cs.DC↗

arXiv2025

IDSS, a Novel P2P Relational Data Storage Service

The rate at which data is generated has been increasing rapidly, raising challenges related to its management. Traditional database management systems suffer from scalability and are usually inefficient when dealing with large-scale and heterogeneous data. This paper introduces IDSS (InnoCyPES Data Storage Service), a novel large-scale data storage tool that leverages peer-to-peer networks and embedded relational databases. We present the IDSS architecture and its design, and provide details related to the implementation. The peer-to-peer framework is used to provide support for distributed queries leveraging a relational database architecture based on a common schema. Furthermore, methods to support complex distributed query processing, enabling robust and efficient management of vast amounts of data are presented.

cs.DB↗

arXiv2025

Accurate Mediterranean Sea forecasting via graph-based deep learning

Accurate ocean forecasting systems are essential for understanding marine dynamics, which play a crucial role in sectors such as shipping, aquaculture, environmental monitoring, and coastal risk management. Traditional numerical solvers, while effective, are computationally expensive and time-consuming. Recent advancements in machine learning have revolutionized weather forecasting, offering fast and energy-efficient alternatives. Building on these advancements, we introduce SeaCast, a neural network designed for high-resolution regional ocean forecasting. SeaCast employs a graph-based framework to effectively handle the complex geometry of ocean grids and integrates external forcing data tailored to the regional ocean context. Our approach is validated through experiments at a high horizontal resolution using the operational numerical forecasting system of the Mediterranean Sea, along with both numerical and data-driven atmospheric forcings. Results demonstrate that SeaCast consistently outperforms the operational model in forecast skill, marking a significant advancement in regional ocean prediction.

physics.ao-ph↗

arXiv2025

Space-Efficient Private Estimation of Quantiles

Fast and accurate estimation of quantiles on data streams coming from communication networks, Internet of Things (IoT), and alike, is at the heart of important data processing applications including statistical analysis, latency monitoring, query optimization for parallel database management systems, and more. Indeed, quantiles are more robust indicators for the underlying distribution, compared to moment-based indicators such as mean and variance. The streaming setting additionally constrains accurate tracking of quantiles, as stream items may arrive at a very high rate and must be processed as quickly as possible and discarded, being their storage usually unfeasible. Since an exact solution is only possible when data are fully stored, the goal in practical contexts is to provide an approximate solution with a provably guaranteed bound on the approximation error committed, while using a minimal amount of space. At the same time, with the increasing amount of personal and sensitive information exchanged, it is essential to design privacy protection techniques to ensure confidentiality and data integrity. In this paper we present the following differentially private streaming algorithms for frugal estimation of a quantile: \textsc{DP-Frugal-1U-L}, \textsc{DP-Frugal-1U-G}, \textsc{DP-Frugal-1U-$ρ$}. Frugality refers to the ability of the algorithms to provide a good approximation to the sought quantile using a modest amount of space, either one or two units of memory. We provide a theoretical analysis and experimental results.

cs.DS↗

arXiv2025

Improving Oil Slick Trajectory Simulations with Bayesian Optimization

Accurate simulations of oil spill trajectories are essential for supporting practitioners' response and mitigating environmental and socioeconomic impacts. Numerical models, such as MEDSLIK-II, simulate advection, dispersion, and transformation processes of oil particles. However, simulations heavily rely on accurate parameter tuning, still based on expert knowledge and manual calibration. To overcome these limitations, we integrate the MEDSLIK-II numerical oil spill model with a Bayesian optimization framework to iteratively estimate the best physical parameter configuration that yields simulation closer to satellite observations of the slick. We focus on key parameters, such as horizontal diffusivity and drift factor, maximizing the Fraction Skill Score (FSS) as a measure of spatio-temporal overlap between simulated and observed oil distributions. We validate the framework for the Baniyas oil incident that occurred in Syria between August 23 and September 4, 2021, which released over 12,000 $m^3$ of oil. We show that, on average, the proposed approach systematically improves the FSS from 5.82% to 11.07% compared to control simulations initialized with default parameters. The optimization results in consistent improvement across multiple time steps, particularly during periods of increased drift variability, demonstrating the robustness of our method in dynamic environmental conditions.

physics.ao-ph↗

arXiv2024

Just In Time Transformers

Precise energy load forecasting in residential households is crucial for mitigating carbon emissions and enhancing energy efficiency; indeed, accurate forecasting enables utility companies and policymakers, who advocate sustainable energy practices, to optimize resource utilization. Moreover, smart meters provide valuable information by allowing for granular insights into consumption patterns. Building upon available smart meter data, our study aims to cluster consumers into distinct groups according to their energy usage behaviours, effectively capturing a diverse spectrum of consumption patterns. Next, we design JITtrans (Just In Time transformer), a novel transformer deep learning model that significantly improves energy consumption forecasting accuracy, with respect to traditional forecasting methods. Extensive experimental results validate our claims using proprietary smart meter data. Our findings highlight the potential of advanced predictive technologies to revolutionize energy management and advance sustainable power systems: the development of efficient and eco-friendly energy solutions critically depends on such technologies.

cs.LG↗

arXiv2021

Data stream fusion for accurate quantile tracking and analysis

UDDSKETCH is a recent algorithm for accurate tracking of quantiles in data streams, derived from the DDSKETCH algorithm. UDDSKETCH provides accuracy guarantees covering the full range of quantiles independently of the input distribution and greatly improves the accuracy with regard to DDSKETCH. In this paper we show how to compress and fuse data streams (or datasets) by using UDDSKETCH data summaries that are fused into a new summary related to the union of the streams (or datasets) processed by the input summaries whilst preserving both the error and size guarantees provided by UDDSKETCH. This property of sketches, known as mergeability, enables parallel and distributed processing. We prove that UDDSKETCH is fully mergeable and introduce a parallel version of UDDSKETCH suitable for message-passing based architectures. We formally prove its correctness and compare it to a parallel version of DDSKETCH, showing through extensive experimental results that our parallel algorithm almost always outperforms the parallel DDSKETCH algorithm with regard to the overall accuracy in determining the quantiles.

cs.DS↗

arXiv2020

UDDSketch: Accurate Tracking of Quantiles in Data Streams

We present UDDSketch (Uniform DDSketch), a novel sketch for fast and accurate tracking of quantiles in data streams. This sketch is heavily inspired by the recently introduced DDSketch, and is based on a novel bucket collapsing procedure that allows overcoming the intrinsic limits of the corresponding DDSketch procedures. Indeed, the DDSketch bucket collapsing procedure does not allow the derivation of formal guarantees on the accuracy of quantile estimation for data which does not follow a sub-exponential distribution. On the contrary, UDDSketch is designed so that accuracy guarantees can be given over the full range of quantiles and for arbitrary distribution in input. Moreover, our algorithm fully exploits the budgeted memory adaptively in order to guarantee the best possible accuracy over the full range of quantiles. Extensive experimental results on synthetic datasets confirm the validity of our approach.

cs.DS↗

arXiv2020

Fast Detection of Outliers in Data Streams with the $Q_n$ Estimator

We present FQN (Fast $Q_n$), a novel algorithm for fast detection of outliers in data streams. The algorithm works in the sliding window model, checking if an item is an outlier by cleverly computing the $Q_n$ scale estimator in the current window. We thoroughly compare our algorithm for online $Q_n$ with the state of the art competing algorithm by Nunkesser et al, and show that FQN (i) is faster, (ii) its computational complexity does not depend on the input distribution and (iii) it requires less space. Extensive experimental results on synthetic datasets confirm the validity of our approach.

cs.DS↗

arXiv2018

Distributed mining of time--faded heavy hitters

We present \textsc{P2PTFHH} (Peer--to--Peer Time--Faded Heavy Hitters) which, to the best of our knowledge, is the first distributed algorithm for mining time--faded heavy hitters on unstructured P2P networks. \textsc{P2PTFHH} is based on the \textsc{FDCMSS} (Forward Decay Count--Min Space-Saving) sequential algorithm, and efficiently exploits an averaging gossip protocol, by merging in each interaction the involved peers' underlying data structures. We formally prove the convergence and correctness properties of our distributed algorithm and show that it is fast and simple to implement. Extensive experimental results confirm that \textsc{P2PTFHH} retains the extreme accuracy and error bound provided by \textsc{FDCMSS} whilst showing excellent scalability. Our contributions are three-fold: (i) we prove that the averaging gossip protocol can be used jointly with our augmented sketch data structure for mining time--faded heavy hitters; (ii) we prove the error bounds on frequency estimation; (iii) we experimentally prove that \textsc{P2PTFHH} is extremely accurate and fast, allowing near real time processing of large datasets.

cs.DC↗

arXiv2018

Mining frequent items in unstructured P2P networks

Large scale decentralized systems, such as P2P, sensor or IoT device networks are becoming increasingly common, and require robust protocols to address the challenges posed by the distribution of data and the large number of peers belonging to the network. In this paper, we deal with the problem of mining frequent items in unstructured P2P networks. This problem, of practical importance, has many useful applications. We design P2PSS, a fully decentralized, gossip--based protocol for frequent items discovery, leveraging the Space-Saving algorithm. We formally prove the correctness and theoretical error bound. Extensive experimental results clearly show that P2PSS provides very good accuracy and scalability, also in the presence of highly dynamic P2P networks with churning. To the best of our knowledge, this is the first gossip--based distributed algorithm providing strong theoretical guarantees for both the Approximate Frequent Items Problem in Unstructured P2P Networks and for the frequency estimation of discovered frequent items.

cs.DS↗

arXiv2017

Fast and Accurate Mining of Correlated Heavy Hitters

The problem of mining Correlated Heavy Hitters (CHH) from a two-dimensional data stream has been introduced recently, and a deterministic algorithm based on the use of the Misra--Gries algorithm has been proposed by Lahiri et al. to solve it. In this paper we present a new counter-based algorithm for tracking CHHs, formally prove its error bounds and correctness and show, through extensive experimental results, that our algorithm outperforms the Misra--Gries based algorithm with regard to accuracy and speed whilst requiring asymptotically much less space.

cs.DS↗