Excited to share our latest quantum chemistry preprint @cqs_lab, the result of the hard work of Clemens Giuliani, arxiv.org/abs/2503.14502. We use a "simple" variational wavefunction composed of a few hundred optimized non-orthogonal Slater determinants and show that it achieves energy accuracies comparable to state-of-the-art methods. While the ansatz itself is as old as quantum chemistry, optimizing it fully has proven challenging so far. The main innovation is an approach to optimize the determinants efficiently, leveraging the quadratic dependence of energy on selected parameters, allowing for exact optimization. ​Using optimized contractions, it has scaling computational cost with the fourth power of the number of basis functions. Benchmarking against exact full-configuration interaction results, we achieved lower variational energies than CCSD(T) for several molecules in the double-zeta basis. It is more accurate than all second-quantized NQS results for molecules so far published, despite being a much simpler ansatz, conceptually. This, in my view, further highlights the fact that for the quantum chemistry of small to intermediate molecules fully correlated wave functions might not be necessarily needed.

Looking for a talented postdoc to join my group @cqs_lab @EPFL_en! 🇨🇭 Research Topics Include: Neural Quantum States, Many-Body Systems, Ab-Initio & Quantum Chemistry...etc Start: Fall 2025, excellent conditions Apply: epfl.ch/labs/cqsl/job-opport…

Our endeavor on variational benchmarks for quantum many-body problems is now published in @ScienceMagazine! In this large collaborative work (~30 institutions) led by @cqs_lab @EPFL_en, we established a practical, agreed-upon metric to quantify the hardness of quantum problems involving many particles (e.g. in materials). The resulting metric, called v-score, is handy to compute and intrinsically suited for variational methods on both classical and quantum devices (e.g. Matrix Product States, Neural Quantum States, Variational Auxiliary Field Monte Carlo, Variational Quantum Eigensolver, DMFT solvers, and many others). One surprising outcome is that this metric almost universally correlates with the error on ground state energies, for problems seemingly unrelated and for the tens of different techniques we have benchmarked. This allows us to identify those problems and regions of interactions and parameters that are factually hard for existing many-body methods (e.g. spin liquids in some 3D geometries, Hubbard models for specific values of the coupling U etc.). The v-score can assess progress of novel computational methods, both classical and quantum-based. It will help shape criteria for measuring quantum computing performance, transitioning from generic, average-hardness qualifiers of computational complexity theory to physics-chemistry based hardness quantifiers for problems central to computational quantum science. In this domain, well-established heuristics exist, making any quantum-driven improvement both significant and highly valued. Read the article here: science.org/doi/10.1126/scie…

Amazing news! Riccardo Rossi of @cqs_lab @EPFL_en has been recognized with the Hermann Kümmel Early Achievement Award for "groundbreaking advances in computational quantum field theory for many-fermion problems, including determinant algorithms for connected-diagram expansions and resummation techniques, leading to key results on the unitary Fermi gas and the Hubbard model". Congratulations Riccardo, so well deserved! actu.epfl.ch/news/riccardo-r…

🔬 New @cqs_lab Publication Alert: Linda Mauron's debut paper, "Predicting Topological Entanglement Entropy in a Rydberg Analog Simulator," marks a significant advancement in the classical simulation of the dynamics of strongly interacting Rydberg atoms in two dimensions. Based on a lightweight, yet accurate variational approach, we can predict entanglement entropies from first principles, for system with more than a few hundred atoms. These simulations, for example, allow to predict a lack of genuine topological entanglement, in experiments performing adiabatic state preparation. Kudos to Linda, @zdenis_, and @JannesNys of @cqs_lab for their contributions! 📄 Preprint here: arxiv.org/abs/2406.19872

New @cqs_lab preprint, with Imelda Romero and @JannesNys ! For those of you interested in computing excited states in periodic systems, check this out arxiv.org/abs/2406.09077, we introduce a neural backflow transformation that allows to accurately target excited states of given momentum. We showcase this on the t-V model of spinless fermions in 2D, where one can nicely characterize excitations through the phase transition from a metal to a charge ordered phase.

Our work introducing neural Pfaffian wave functions to describe pairing in ultra-cold Fermi gases has been published here rdcu.be/dHM9w ! Also see this very nice blog post go.nature.com/3Ua8GF6 by Jane Kim, Alessandro Lovato & Bryce Fore

The time-dependent Schrodinger equation is the building block of non-relativistic quantum theory. Yet, solving it for electronic systems in continuous space, in regimes where correlations cannot be neglected, is a very hard theoretical and computational challenge, at least as old as quantum mechanics itself. In this last @cqs_lab preprint with @JannesNys and @pgabbo0 we show how time-dependent, correlated electronic wave functions can be used to track the dynamics in regimes where td hartree fock gives qualitatively wrong results. With time-dependent variational Monte Carlo, we use Jastrow and Jastrow-backflow variational states, also considering neural network parameterizations. The video below here (courtesy of @JannesNys) shows the density distribution in a H2 molecule under an intense laser field, one of the applications we consider in the manuscript, together with quantum dots and more! Check it out here arxiv.org/abs/2403.07447

We had a very productive @cqs_lab group retreat last week, enjoying the beauty of southern Italy (Greek temples, pizzas, and buffaloes included) and discussing quantum physics, @NetKetOrg and more!