(back to the home, torna alla home)
My early career research achievements can be synthesised as follows:
Exact results in the quantum many-body problem and quantum information. Mainly during my PhD (2007-2011, supervised by Prof. Germán Sierra) I contributed to three corpora of results in theoretical physics: the study of the quantum entanglement (of elementary excitations) in low-dimensional many-body systems; the exact solution of a family of pairing models of superfluidity and superconductivity; the analysis of quantum entanglement among non-spatial partitions of the Hilbert space (energy and momentum space) in extended quantum systems.
We demonstrated, in particular, exact (and celebrated, see [Goldstein+ 2018]) laws relating the entanglement entropy of quantum excitations in low-dimensional critical ensembles of interacting bodies, to the universal quantities of the field theories that describe their continuum limit.
I these projects, I derived part of the analytical results (as the entanglement entropy of some of the primary operators in [IB+ 2010]), and developed the algorithms (mainly exact diagonalization, integration of Bethe ansatz and BCS equations in different ensembles) for the numerical solution of the Schroedinger eigenvalue problem and for the computation of the entanglement entropy of the associated wavefunctions.
Fundamental statistical mechanics of disordered systems. In two subsequent post-doctoral fellowships (IPCF-CNR/Sapienza University, 2011-2014; INFN-Parma/University of Parma, 2014-2016) I have worked in several fundamental topics in classical statistical mechanics. Mainly: the emergence of chaos in biologically inspired models of neural dynamics; the study of critical behaviour and the superfluid transition in vector statistical models on disordered and long-range networks; the statistical physical description of nonlinear processes in lasers and random lasers.
We successfully modelled, in particular, the transition to laser generation as a thermodynamic phase transition, uncovering the various emergent properties of the relevant (XY, spherical, complex-spherical) spin models, such as ageing, replica symmetry breaking, non-equipartite energy condensation, and 1st-order/continuous/topological phase transitions, as a function of the sparsity and disorder of the network of (pairwise and 4th-order) interactions.
In these years, I developed as well a fundamental description of finite size effects in metastable phases of some statistical models.
On the methodological side, I developed ad hoc, original algorithms for (parallel and high-performance, see [IB+ 2013]) Markov Chain Monte Carlo sampling of vector spin models on disordered networks and networks of quadruplet couplings (see [Antenucci+ 2015]). I also applied the so called Langer theory of metastability [Binder 1987], as well as the evaporation-condensation theory, beyond the well-known Ising model paradigm, validating both against numerical (cluster, Wolff-Swendsen-Wang) dynamic MCMC sampling.
Statistical inference applied to neuroscience and cognition/decision making. During the second research internship in Sapienza University (2017-2020), I happened to develop a novel experimental scheme of analysis of the phenomenon of facial attractiveness. By using an interactive data-generation method where participants “sculpted” their preferred facial images, along with unsupervised probabilistic models, we demonstrated, with unprecedented precision (to the best of my knowledge), that facial attractiveness is fundamentally subjective, and strongly determined by observer’s gender.
On the methodological side, I devised the experimental design and develop the associated software for the human-computer interaction algorithm (based on image deformation and genetic algorithms). I developed the tools for the statistical analysis (based on maximum entropy models and Boltzmann learning, see [IB+ 2019], [IB+ 2020]) devoted to the unsupervised inference of the subject’s gender, and to the estimation of the relevant order of the interaction between facial landmarks.
In this period, I reviewed and developed as well statistical methods of noise-cleaning of correlation and precision matrices, based on Random Matrix Theory, and applied them to the analysis of human brain fMRI activity series —a line of research that I keep working on.
Analytical theory of non-Markovian stochastic processes, with applications to information transmission in neural networks. During my internship in IIT (2021-2023), I developed a theory of information transmission in neural networks, describing how (pre-synaptic) neuronal noise correlations shape information transmission in (post-synaptic) biological and artificial neural networks. In particular, we successfully provided a clear theoretical explanation of a seemingly paradoxical phenomenon regarding the controversial influence of noise correlations in brain neural populations [IB+ 2025], according to which information-limiting noise correlations, that are detrimental to decode the baseline activity of a neural population, may enhance the information of such baseline activity that is transmitted to a decoding, readout population. This phenomenon [Valente+ 2021] could play an important role in how information is transmitted in neural populations, explaining the proliferation of information-limiting noise correlations that have been observed in real neural codes.
On the methodological side, I derived (through a simple separation of timescales ansatz) one of the (apparently [Lindner 2009]) few analytical results regarding the first passage time statistics of a fundamental non-Markovian stochastic processes: the colored-noise random walker in a harmonic potential.