arXiv subjects

Lorenzo Stella

Publications and source records attributed to Lorenzo Stella.

At least 19 recordsLinked to original sources

arXiv2026

Ultrafast Nonthermal Lattice Destabilization and Suppression of Polar Optical Scattering in Electronically Excited $α$-SiO$_2$ from First-Principles and Deep Neural Network Potential Modeling

We present a multiscale first-principles-to-machine-learning approach to investigate ultrafast lattice dynamics in electronically excited $α$-SiO$_2$. Ab initio molecular dynamics (AIMD) based on electronic-temperature-dependent density functional theory (DFT) are used to train electronic-temperature-dependent deep neural network potentials (DNNPs). The use of DNNPs enables atomistic modeling at near-DFT accuracy of large $α$-SiO$_2$ cells with thousands of atoms. In particular, DNNPs allowed us to obtain accurate phonon band structures and molecular dynamics (MD) of $α$-SiO$_2$ excited by a sudden increase in electronic temperature. With increasing electronic temperature, $T_e$, pronounced lattice destabilization of $α$-SiO$_2$ is found, as evidenced by violations of elastic stability criteria, substantial volumetric expansion, a sharp reduction of the bulk modulus, and progressive weakening of Si-O bonding due to antibonding-state occupation. From the electronic and phonon band structures, we estimated the Frohlich coupling constant, which decreases as $T_e$ increases, suggesting a crossover to a nonpolar phase of $α$-SiO$_2$ at elevated electronic temperature. This is corroborated by the Bader charge analysis. We also suggest that polar optical phonon scattering should be strongly suppressed at $T_e > 2$ eV. From large-cell DNNP-MD simulations, we show that a well-defined thermal equilibrium, as defined by the Maxwell-Boltzmann distribution, is not achieved over the first few hundred femtoseconds. This behavior explains the non-monotonic equilibration of the kinetic temperature after a sudden rise of $T_e$. After $T_e$ is raised to 2.6 eV, Si and O atoms first equilibrate separately at two different temperatures, suggesting an atomic fluid phase, in agreement with recent experimental and theoretical findings.

cond-mat.mtrl-sci↗

arXiv2026

fev-bench: A Realistic Benchmark for Time Series Forecasting

Benchmark quality is critical for meaningful evaluation and sustained progress in time series forecasting, particularly with the rise of pretrained models. Existing benchmarks often have limited domain coverage or overlook real-world settings such as tasks with covariates. Their aggregation procedures frequently lack statistical rigor, making it unclear whether observed performance differences reflect true improvements or random variation. Many benchmarks lack consistent evaluation infrastructure or are too rigid for integration into existing pipelines. To address these gaps, we propose fev-bench, a benchmark of 100 forecasting tasks across seven domains, including 46 with covariates. Supporting the benchmark, we introduce fev, a lightweight Python library for forecasting evaluation emphasizing reproducibility and integration with existing workflows. Using fev, fev-bench employs principled aggregation with bootstrapped confidence intervals to report performance along two dimensions: win rates and skill scores. We report results on fev-bench for pretrained, statistical, and baseline models and identify promising future research directions.

cs.LG↗

arXiv2026

β-Ga2O3-Based Heterojunctions: Exploring Growth Orientations and Alloying on Electronic Properties

We investigate the effects of alloying and growth orientation on the electronic properties of the ultra-wide bandgap semiconductor β-Ga2O3 and pseudomorphic (AlxGa1-x)2O3 alloy heterojunctions. Band offsets are computed from first principles using density functional theory (DFT) with the Heyd-Scuseria-Ernzerhof hybrid functional for different Al concentrations and four growth orientations, namely (100)B, (010), (001)B, and ($\bar{2}$01). Significant variations are found and ascribed to the strained pseudomorphic alloys. The values of the band offsets are fed into technology computer-aided design (TCAD) models of Schottky barrier diodes (SBD). I-V and C-V characteristics from the TCAD models show reasonable agreement with recent experimental measurements in the forward bias region. Discrepancies in the negative bias region are expected due to the ideality of the Schottky junctions considered in this study. Our findings underscore the critical role of growth orientation and strain in the accurate modelling of β-Ga2O3-based SBD.

cond-mat.mtrl-sci↗

arXiv2026

Universal Stability of Ga Split Vacancies across α-, β-, and κ-Ga2O3 Polymorphs: A Machine-Learning Accelerated Study

Split Ga vacancies are the dominant native acceptor in $β$-$Ga_2O_3$; however, their role in $α$ and $κ$ phases has been largely overlooked or assumed to be unfavorable. A detailed understanding of these defects is critical for tailoring the electrical conductivity and optical properties and optimising $Ga_2O_3$-based devices. In this work, we used machine learning interatomic potentials (MLIPs) to accelerate the discovery of non-local defect reconstructions, followed by HSE06 hybrid DFT to accurately quantify defect properties of single vacancy $V_{\text{Ga}}$, split vacancy $V_{\text{Ga}}^{\text{i}}$ and substitutional donors ($\mathrm{Hf_{Ga}}$ and $\mathrm{Si_{Ga}}$) across a wide range of experimentally relevant conditions for the oxygen chemical potential. We find that split vacancies are the ground-state vacancy for all studied polymorphs ($β$, $α$, and $κ$). Split vacancies are more stable than simple vacancies by ~0.75 eV ($β$), ~0.41 eV ($α$), and ~0.14 eV ($κ$). Notably, MLIPs correctly identified the specific split-vacancy ground states and yielded an energetic ordering of symmetry-inequivalent defect configurations in excellent agreement with HSE06 results. While Hf and Si show low formation energy and act as shallow donors, especially under oxygen-poor conditions, their efficiency is limited by split-vacancy compensation. The growth under oxygen-poor conditions is a universal requirement to suppress these defects and achieve high n-type conductivity across the $Ga_2O_3$ polymorph.

cond-mat.mtrl-sci↗

arXiv2025

Chronos-2: From Univariate to Universal Forecasting

Pretrained time series models have enabled inference-only forecasting systems that produce accurate predictions without task-specific training. However, existing approaches largely focus on univariate forecasting, limiting their applicability in real-world scenarios where multivariate data and covariates play a crucial role. We present Chronos-2, a pretrained model capable of handling univariate, multivariate, and covariate-informed forecasting tasks in a zero-shot manner. Chronos-2 employs a group attention mechanism that facilitates in-context learning (ICL) through efficient information sharing across multiple time series within a group, which may represent sets of related series, variates of a multivariate series, or targets and covariates in a forecasting task. These general capabilities are achieved through training on synthetic datasets that impose diverse multivariate structures on univariate series. Chronos-2 delivers state-of-the-art performance across three comprehensive benchmarks: fev-bench, GIFT-Eval, and Chronos Benchmark II. On fev-bench, which emphasizes multivariate and covariate-informed forecasting, Chronos-2's universal ICL capabilities lead to substantial improvements over existing models. On tasks involving covariates, it consistently outperforms baselines by a wide margin. Case studies in the energy and retail domains further highlight its practical advantages. The in-context learning capabilities of Chronos-2 establish it as a general-purpose forecasting model that can be used "as is" in real-world forecasting pipelines.

cs.LG↗

arXiv2025

Tailoring the Electronic Properties of Monoclinic (InxAl1-x)2O3 Alloys via Substitutional Donors and Acceptors

Ultra-wide bandgap semiconductors such as \b{eta}-Ga2O3 are ideal materials for next-generation power electronic devices. Electronic and mechanical properties of \b{eta}-Ga2O3 can be tuned by alloying with other sesquioxides, notably Al2O3 and In2O3. Moreover, by tuning the In content of a (InxAl1-x)2O3 alloy, its lattice constants can be matched to those of Ga2O3, while preserving a large conduction-band offset. In view of potential applications to \b{eta}-Ga2O3-based heterostructure, we performed atomistic modelling of (InxAl1-x)2O3 alloys using density functional theory to investigate thermodynamic and electrical properties of conventional group IV dopants (Si, Sn, C, Ge), alternative metal donors (Ta, Zr, Hf), and acceptors (Mg, Zn, Cu). The hybrid Heyd-Scuseria-Ernzerhof functional (HSE06) is used to accurately quantify the defect formation energies, ionization levels, and concentrations over a wide range of experimentally relevant conditions for the oxygen chemical potential and temperature. In our atomistic models, Hf and Zr show favourable properties as alternative donors to Si and other group IV impurities, especially under oxygen-poor conditions. Our findings also suggest that acceptors Mg, Zn, and Cu, while they cannot promote p-doping, can be still beneficial for the compensation of unintentionally n-doped materials, e.g., to generate semi-insulating layers and improve rectification.

cond-mat.mtrl-sci↗

arXiv2025

Zero-Shot Time Series Forecasting with Covariates via In-Context Learning

Pretrained time series models, capable of zero-shot forecasting, have demonstrated significant potential in enhancing both the performance and accessibility of time series forecasting. However, existing pretrained models either do not support covariates or fail to incorporate them effectively. We introduce COSMIC, a zero-shot forecasting model that utilizes covariates via in-context learning. To address the challenge of data scarcity, we propose Informative Covariate Augmentation, which enables the training of COSMIC without requiring any datasets that include covariates. COSMIC achieves state-of-the-art performance in zero-shot forecasting, both with and without covariates. Our quantitative and qualitative analysis demonstrates that COSMIC effectively leverages covariates in zero-shot forecasting.

cs.LG↗

arXiv2025

ChronosX: Adapting Pretrained Time Series Models with Exogenous Variables

Covariates provide valuable information on external factors that influence time series and are critical in many real-world time series forecasting tasks. For example, in retail, covariates may indicate promotions or peak dates such as holiday seasons that heavily influence demand forecasts. Recent advances in pretraining large language model architectures for time series forecasting have led to highly accurate forecasters. However, the majority of these models do not readily use covariates as they are often specific to a certain task or domain. This paper introduces a new method to incorporate covariates into pretrained time series forecasting models. Our proposed approach incorporates covariate information into pretrained forecasting models through modular blocks that inject past and future covariate information, without necessarily modifying the pretrained model in consideration. In order to evaluate our approach, we introduce a benchmark composed of 32 different synthetic datasets with varying dynamics to evaluate the effectivity of forecasting models with covariates. Extensive evaluations on both synthetic and real datasets show that our approach effectively incorporates covariate information into pretrained models, outperforming existing baselines.

cs.LG↗

arXiv2024

Chronos: Learning the Language of Time Series

We introduce Chronos, a simple yet effective framework for pretrained probabilistic time series models. Chronos tokenizes time series values using scaling and quantization into a fixed vocabulary and trains existing transformer-based language model architectures on these tokenized time series via the cross-entropy loss. We pretrained Chronos models based on the T5 family (ranging from 20M to 710M parameters) on a large collection of publicly available datasets, complemented by a synthetic dataset that we generated via Gaussian processes to improve generalization. In a comprehensive benchmark consisting of 42 datasets, and comprising both classical local models and deep learning methods, we show that Chronos models: (a) significantly outperform other methods on datasets that were part of the training corpus; and (b) have comparable and occasionally superior zero-shot performance on new datasets, relative to methods that were trained specifically on them. Our results demonstrate that Chronos models can leverage time series data from diverse domains to improve zero-shot accuracy on unseen forecasting tasks, positioning pretrained models as a viable tool to greatly simplify forecasting pipelines.

cs.LG↗

arXiv2024

Adaptive proximal algorithms for convex optimization under local Lipschitz continuity of the gradient

Backtracking linesearch is the de facto approach for minimizing continuously differentiable functions with locally Lipschitz gradient. In recent years, it has been shown that in the convex setting it is possible to avoid linesearch altogether, and to allow the stepsize to adapt based on a local smoothness estimate without any backtracks or evaluations of the function value. In this work we propose an adaptive proximal gradient method, adaPG, that uses novel estimates of the local smoothness modulus which leads to less conservative stepsize updates and that can additionally cope with nonsmooth terms. This idea is extended to the primal-dual setting where an adaptive three-term primal-dual algorithm, adaPD, is proposed which can be viewed as an extension of the PDHG method. Moreover, in this setting the "essentially" fully adaptive variant adaPD$^+$ is proposed that avoids evaluating the linear operator norm by invoking a backtracking procedure, that, remarkably, does not require extra gradient evaluations. Numerical simulations demonstrate the effectiveness of the proposed algorithms compared to the state of the art.

math.OC↗

arXiv2023

Deep Non-Parametric Time Series Forecaster

This paper presents non-parametric baseline models for time series forecasting. Unlike classical forecasting models, the proposed approach does not assume any parametric form for the predictive distribution and instead generates predictions by sampling from the empirical distribution according to a tunable strategy. By virtue of this, the model is always able to produce reasonable forecasts (i.e., predictions within the observed data range) without fail unlike classical models that suffer from numerical stability on some data distributions. Moreover, we develop a global version of the proposed method that automatically learns the sampling strategy by exploiting the information across multiple related time series. The empirical evaluation shows that the proposed methods have reasonable and consistent performance across all datasets, proving them to be strong baselines to be considered in one's forecasting toolbox.

cs.LG↗

arXiv2022

Thermoelectric properties of cement composite analogues from first principles calculations

Buildings are responsible for a considerable fraction of the energy wasted globally every year, and as a result, excess carbon emissions. While heat is lost directly in colder months and climates, resulting in increased heating loads, in hot climates cooling and ventilation is required. One avenue towards improving the energy efficiency of buildings is to integrate thermoelectric devices and materials within the fabric of the building to exploit the temperature gradient between the inside and outside to do useful work. Cement-based materials are ubiquitous in modern buildings and present an interesting opportunity to be functionalised. We present a systematic investigation of the electronic transport coefficients relevant to the thermoelectric materials of the calcium silicate hydrate (C-S-H) gel analogue, tobermorite, using Density Functional Theory calculations with the Boltzmann transport method. The calculated values of the Seebeck coefficient are within the typical magnitude (200 - 600 $μV/K$) indicative of a good thermoelectric material. The tobermorite models are predicted to be intrinsically $p$-type thermoelectric material because of the presence of large concentration of the Si-O tetrahedra sites. The calculated electronic $ZT$ for the tobermorite models have their optimal values of 0.983 at (400 $\mathrm{K}$ and $10^{17}$ $\mathrm{cm^{-3}}$) for tobermorite 9 Å, 0.985 at (400 $\mathrm{K}$ and $10^{17}$ $\mathrm{cm^{-3}}$) for tobermorite 11 Å and 1.20 at (225 $\mathrm{K}$ and $10^{19}$ $\mathrm{cm^{-3}}$) for tobermorite 14 Å, respectively.

cond-mat.mtrl-sci↗

arXiv2022

Deep Learning for Time Series Forecasting: Tutorial and Literature Survey

Deep learning based forecasting methods have become the methods of choice in many applications of time series prediction or forecasting often outperforming other approaches. Consequently, over the last years, these methods are now ubiquitous in large-scale industrial forecasting applications and have consistently ranked among the best entries in forecasting competitions (e.g., M4 and M5). This practical success has further increased the academic interest to understand and improve deep forecasting methods. In this article we provide an introduction and overview of the field: We present important building blocks for deep forecasting in some depth; using these building blocks, we then survey the breadth of the recent deep forecasting literature.

cs.LG↗

arXiv2021

Douglas-Rachford splitting and ADMM for nonconvex optimization: Accelerated and Newton-type linesearch algorithms

Although the performance of popular optimization algorithms such as Douglas-Rachford splitting (DRS) and the ADMM is satisfactory in small and well-scaled problems, ill conditioning and problem size pose a severe obstacle to their reliable employment. Expanding on recent convergence results for DRS and ADMM applied to nonconvex problems, we propose two linesearch algorithms to enhance and robustify these methods by means of quasi-Newton directions. The proposed algorithms are suited for nonconvex problems, require the same black-box oracle of DRS and ADMM, and maintain their (subsequential) convergence properties. Numerical evidence shows that the employment of L-BFGS in the proposed framework greatly improves convergence of DRS and ADMM, making them robust to ill conditioning. Under regularity and nondegeneracy assumptions at the limit point, superlinear convergence is shown when quasi-Newton Broyden directions are adopted.

math.OC↗

arXiv2020

Anomaly Detection at Scale: The Case for Deep Distributional Time Series Models

This paper introduces a new methodology for detecting anomalies in time series data, with a primary application to monitoring the health of (micro-) services and cloud resources. The main novelty in our approach is that instead of modeling time series consisting of real values or vectors of real values, we model time series of probability distributions over real values (or vectors). This extension to time series of probability distributions allows the technique to be applied to the common scenario where the data is generated by requests coming in to a service, which is then aggregated at a fixed temporal frequency. Our method is amenable to streaming anomaly detection and scales to monitoring for anomalies on millions of time series. We show the superior accuracy of our method on synthetic and public real-world data. On the Yahoo Webscope data set, we outperform the state of the art in 3 out of 4 data sets and we show that we outperform popular open-source anomaly detection tools by up to 17% average improvement for a real-world data set.

cs.LG↗

arXiv2020

Proximal Gradient Algorithms: Applications in Signal Processing

Advances in numerical optimization have supported breakthroughs in several areas of signal processing. This paper focuses on the recent enhanced variants of the proximal gradient numerical optimization algorithm, which combine quasi-Newton methods with forward-adjoint oracles to tackle large-scale problems and reduce the computational burden of many applications. These proximal gradient algorithms are here described in an easy-to-understand way, illustrating how they are able to address a wide variety of problems arising in signal processing. A new high-level modeling language is presented which is used to demonstrate the versatility of the presented algorithms in a series of signal processing application examples such as sparse deconvolution, total variation denoising, audio de-clipping and others.

eess.SP↗

arXiv2019

GluonTS: Probabilistic Time Series Models in Python

We introduce Gluon Time Series (GluonTS, available at https://gluon-ts.mxnet.io), a library for deep-learning-based time series modeling. GluonTS simplifies the development of and experimentation with time series models for common tasks such as forecasting or anomaly detection. It provides all necessary components and tools that scientists need for quickly building new models, for efficiently running and analyzing experiments and for evaluating model accuracy.

cs.LG↗

arXiv2018

Newton-type Alternating Minimization Algorithm for Convex Optimization

We propose NAMA (Newton-type Alternating Minimization Algorithm) for solving structured nonsmooth convex optimization problems where the sum of two functions is to be minimized, one being strongly convex and the other composed with a linear mapping. The proposed algorithm is a line-search method over a continuous, real-valued, exact penalty function for the corresponding dual problem, which is computed by evaluating the augmented Lagrangian at the primal points obtained by alternating minimizations. As a consequence, NAMA relies on exactly the same computations as the classical alternating minimization algorithm (AMA), also known as the dual proximal gradient method. Under standard assumptions the proposed algorithm possesses strong convergence properties, while under mild additional assumptions the asymptotic convergence is superlinear, provided that the search directions are chosen according to quasi-Newton formulas. Due to its simplicity, the proposed method is well suited for embedded applications and large-scale problems. Experiments show that using limited-memory directions in NAMA greatly improves the convergence speed over AMA and its accelerated variant.

math.OC↗