Lorenzo Dall'Amico

Lorenzo Dall'Amico

PhD student

Gipsa lab, Grenoble INP

Biography

I am a PhD studen in the GAIA team of GIPSA lab in Grenoble, under the supervision of Prof. Romain Couillet and of Nicolas Tremblay with a project meant to improve the state of the art knowledge of theoretical machine learning.

Interests

  • Statistical physics
  • Machine learning

Education

  • Master degree in Physics of complex systems, 2018

    Politecnico di Torino

  • M2 in Physics of complex systems, 2018

    Paris Sud (XI)

  • BSc in physical engineering, 2016

    Politecnico di Torino

CoDeBetHe

Community detection with the Bethe Hessian

Find here an efficient Julia implementation of our algorithms for community detection

Get the packge View Documentation

Projects

Limit order book models

Can we exploit statistical physics of complex systems to model the behavior of the financial market?.

Spectral clustering in sparse graphs

Retrivieng the community structure of a sparse graph is a challenging and deeply studied problem. See here our contributions.

Recent Publications

Quickly discover relevant content by filtering publications.

Nishimori meets Bethe: a spectral method for node classification in sparse weighted graphs

In this article we show a relation between the Bethe approximation for Ising model on random graphs and the Nishimori temperature with applications to unsupervised correlation clustering

Community detection in sparse time-evolving graphs with a dynamical Bethe-Hessian

In this article we propose an new spectral algorithm to perform spectral clustering in sparse, dynamical graphs.

Optimal Laplacian regularization for sparse spectral community detection

In this article we provide a new interpretation on why regularization helps in sparse community detection as well as providing results on the optimal regularization to be adopted.

A unified framework for spectral clustering in sparse graphs

In this article we show how the benchmark spectral clustering techniques can all be explained under the same, elegant framework. We further provide a highly performing algorithm for community detection.

Revisiting the Bethe-Hessian: Improved Community Detection in Sparse Heterogeneous Graphs

In this article we propose an improved regularization of the Bethe-Hessian matrix to perform spectral clustering in sparse graphs.

Contact