arXiv subjects

Massimo Bortone

Publications and source records attributed to Massimo Bortone.

5 recordsLinked to original sources

arXiv2026

Machine Learning Interatomic Potentials: Advancing Open-Source Software for Efficient and Scalable Molecular Simulation

Machine learning interatomic potentials (MLIPs) enable atomistic simulations with near ab initio accuracy at significantly reduced computational cost, but their broader adoption is often limited by fragmented tooling, limited scalability, and inflexible software design. We present mlip v2, a new generation of the mlip library that advances efficient and scalable molecular simulation through a unified and extensible framework. The new release features a targeted API redesign with improved modularity and control, enabling flexible customization of training, data processing, and simulation workflows. It further integrates a new high-performance backend for equivariant operations, e3j, significantly accelerating model inference and simulations. In addition, the framework introduces a range of entirely new capabilities, including the eSEN architecture with a Mixture-of-Experts formulation for scalable training on large and diverse datasets, improved handling of electrostatics through more physically grounded charge modeling and long-range interaction treatment, and advanced simulation features such as NPT ensembles and nudged elastic band methods. Together, these extensions significantly broaden the scope of MLIP applications, enabling efficient modeling of complex, reactive, and out-of-equilibrium systems, and bridging the gap between ML research and practical molecular simulation applications. The library is available on GitHub and on PyPI under the Apache license 2.0.

physics.chem-ph↗

arXiv2025

Bayesian perspectives for quantum states and application to ab initio quantum chemistry

The quantum many-electron problem is not just at the heart of condensed matter phenomena, but also essential for first-principles simulation of chemical phenomena. Strong correlation in chemical systems are prevalent and present a formidable challenge in the simulation of these systems, while predictive phenomena in this domain often also requires a demanding level of accuracy to inform chemical behavior. Efficient representations of the many-electron states of chemical systems are therefore also being inspired by machine learning principles to provide an alternative to established approaches. In this chapter, we review recent progress in this endeavor for quantum chemical problems represented in second quantization, and the particular challenges present in this field. In particular, we focus on the application of Gaussian Process States emerging from efficient representations of the many-body wavefunction with rigorous Bayesian modeling frameworks, allowing for the unification of multiple paradigms under a common umbrella. We show how such models (and other representations derived from machine learning) can be used as novel tools to compute ab initio chemical properties, while in turn also informing the design of machine learning models to extract correlation patterns in classical data.

cond-mat.str-el↗

arXiv2024

Simple Fermionic backflow states via a systematically improvable tensor decomposition

We present an effective ansatz for the wave function of correlated electrons that brings closer the fields of machine learning parameterizations and tensor rank decompositions. We consider a CANDECOMP/PARAFAC (CP) tensor factorization of a general backflow transformation in second quantization for a simple, compact and systematically improvable Fermionic state. This directly encodes $N$-body correlations without the ordering dependence of other tensor decompositions. We consider and explicitly demonstrate various controllable truncations, in the rank and range of the backflow correlations or magnitude of local energy contributions, in order to systematically affect scaling reductions to $\mathcal{O}[N^{3-4}]$. Benchmarking against small Fermi-Hubbard and chemical systems reveals an improvement over other NQS-like models, while extending towards larger strongly correlated ab initio systems demonstrates competitive accuracy with more established DMRG techniques on ab initio 2D hydrogenic lattices with realistic long-range Coulomb interactions.

quant-ph↗

arXiv2023

Impact of conditional modelling for a universal autoregressive quantum state

We present a generalized framework to adapt universal quantum state approximators, enabling them to satisfy rigorous normalization and autoregressive properties. We also introduce filters as analogues to convolutional layers in neural networks to incorporate translationally symmetrized correlations in arbitrary quantum states. By applying this framework to the Gaussian process state, we enforce autoregressive and/or filter properties, analyzing the impact of the resulting inductive biases on variational flexibility, symmetries, and conserved quantities. In doing so we bring together different autoregressive states under a unified framework for machine learning-inspired ans\"atze. Our results provide insights into how the autoregressive construction influences the ability of a variational model to describe correlations in spin and fermionic lattice models, as well as ab initio electronic structure problems where the choice of representation affects accuracy. We conclude that, while enabling efficient and direct sampling, thus avoiding autocorrelation and loss of ergodicity issues in Metropolis sampling, the autoregressive construction materially constrains the expressivity of the model in many systems.

quant-ph↗

arXiv2020

Optimizing the energy consumption of spiking neural networks for neuromorphic applications

In the last few years, spiking neural networks have been demonstrated to perform on par with regular convolutional neural networks. Several works have proposed methods to convert a pre-trained CNN to a Spiking CNN without a significant sacrifice of performance. We demonstrate first that quantization-aware training of CNNs leads to better accuracy in SNNs. One of the benefits of converting CNNs to spiking CNNs is to leverage the sparse computation of SNNs and consequently perform equivalent computation at a lower energy consumption. Here we propose an efficient optimization strategy to train spiking networks at lower energy consumption, while maintaining similar accuracy levels. We demonstrate results on the MNIST-DVS and CIFAR-10 datasets.

cs.NE↗