Science

Time Series Analysis in Machine Learning

Chapter for the book "Machine Learning Techniques for Astrophysics and Cosmology" (Eds. Cosimo Bambi, Vinay Kashyap, Swarnim Shashank, Naoki Yoshida — Springer Singapore, expected 2027). Review of time series analysis from a machine learning perspective: from classical statistical models (ARIMA, exponential smoothing, state-space) to modern methods — tree ensembles, hidden Markov models, Gaussian processes, and deep learning (RNNs, CNNs, transformers).

Machine Learning for Event Reconstruction in Imaging Atmospheric Cherenkov Telescopes

Chapter for the book "Machine Learning Techniques for Astrophysics and Cosmology" (Eds. Cosimo Bambi, Vinay Kashyap, Swarnim Shashank, Naoki Yoshida — Springer Singapore, expected 2027). Review of the role of machine learning in event reconstruction at Imaging Atmospheric Cherenkov Telescopes (IACTs): from the classical Hillas-parameter pipeline to timing-based features, ensemble methods (gradient boosting, stacking), and Convolutional/Graph Neural Networks.

Application of Machine and Deep Learning Methods to the Analysis of IACTs Data

Book chapter (Springer, ISBN 978-3-030-65867-0). Gamma/hadron discrimination and muon tagging in Imaging Atmospheric Cherenkov Telescope data using ML and Deep Learning, with applications to optical calibration and cosmic-ray background rejection.

Machine Learning-Enhanced Discrimination of Gamma-Ray and Hadron Events Using Temporal Features: An ASTRI Mini-Array Analysis

A Random Forest method to improve gamma-ray vs. cosmic-ray discrimination in ASTRI Mini-Array Cherenkov telescopes, leveraging pixel temporal information beyond traditional morphological features.

Application of Machine Learning Ensemble Methods to ASTRI Mini-Array Cherenkov Event Reconstruction

Gamma/hadron separation and energy reconstruction in ASTRI Mini-Array Cherenkov telescopes: a Stacking ensemble of Extra Trees, Random Forest and XGBoost proves the most sensitive ML technique for signal segregation.

Regime-Aware LightGBM for Stock Market Forecasting: A Validated Walk-Forward Framework with Statistical Rigor and Explainable AI Analysis

A stock market forecasting framework based on regime-aware LightGBM conditioned on Hidden Markov Model market regimes, tested on 51 NASDAQ-100 stocks with a Sharpe ratio of 1.18.

Cognitive Biases in Asset Pricing: An Empirical Analysis of the Alphabet Effect and Ticker Fluency in the US Market

Empirical analysis of cognitive biases related to ticker processing fluency in the S&P 500: in the era of algorithmic trading, the alphabet effect shows no statistically significant impact on returns.

Artificial Intelligence vs. Efficient Markets: A Critical Reassessment of Predictive Models in the Big Data Era

Critical review of AI applications in financial market prediction: ensemble methods systematically outperform single classifiers, yet many statistically significant models fail to generate real economic value net of transaction costs.

Forecasting Significant Stock Market Price Changes Using Machine Learning: Extra Trees Classifier Leads

An Extra Trees Classifier model for predicting significant 10-day stock price changes, trained on technical indicators across 120 companies, achieving 86.1% accuracy.

An Introduction to Machine Learning Methods for Fraud Detection

A review of Machine Learning techniques for financial fraud detection (credit cards, financial statements, insurance, money laundering), with two case studies on real banking data.

Quantitative measurement of cognitive biases in LLMs using the methodology of experimental physics: repeated measurements with uncertainty decomposed into statistical and systematic components. Introduces the Bias Strength Index (BSI), evaluates eleven biases across eight LLMs on over 70,000 responses, and analyzes three inference-time debiasing strategies.

A review of ML and Deep Learning methods for cloud masking in multispectral satellite imagery, with emphasis on DL approaches that improve accuracy and efficiency in cloudy pixel classification.

Guest editorial for a Special Issue on advanced AI techniques for time series analysis and pattern recognition, spanning data-driven finance, astrophysical event reconstruction, and beyond.

An overview of AI integration in scientific experimentation, from physics to biology, tracing the evolution from the 1990s to future prospects with deep learning and generative models.

Analysis of the growing specialization of AI across scientific and industrial domains: progress is measured by the co-design of models with domain-specific knowledge to solve high-impact real-world problems.