






Welcome
to my homepage
Here
you'll find information about my activities at the university and
out of it


You
can always reach me by email at Ferran.Mazzanti@upc.edu 



My main fields
of research are Condensed Matter and Quantum ManyBody physics,
although I also work on a very specific kind on neural networks
called Boltzmann machines. More specifically my preferred topics
are







Liquid
Helium
Helium
is the lightest of the noble gases and due to its low mass
shows amazing quantum properties that can be observed at amacroscopic
level, such as BoseEinstein Condensation and Superfluidity.Besides,
the interatomic potential between pairs of Helium atoms is
fairly well known and one can easily calculate static
and (some) dynamic properties at zero and low temeratures
using either variational or Monte carlo techniques.Superfluid
Helium was discovered in the beginning of the 20th century
and has since then attracted the interest of many nowadays
well honored physicists. Despite Helium has been studied for
more than fifty years, there are still some aspects not very
well understood such as the exact relation between a Bose
condensate and superfluidity.Homogeneous
and inhomogeneous Helium, 1D, 2D or 3D Helium, liquid, solid,
pure or in mixtures... what's your favourie flavour?

BoseEinstein
condensation
BoseEinstein
Condensation (BEC) is perhaps the most uptothedate field
of interest in manybody physics. it was as late as 1995 when
Eric Cornell and Carl Wieman in Boulder, Colorado, achieved
the first neat realization of a Bose condensate conatining almost
100% of particles on it. Up to that date only indirect meaurements
of Bose condensates where performed, as for instance the extra
strength in the peak of the dynamic structure function of 4He
at high momentum transfer. Many things have happened since 1995
and nowadays experimental physicists are able to do wonders
with these magnetically trapped condensates, and the number
of atoms confined is increasing every day.

Dinyamic
Structure Function of Quantum Fluids
One
of my main interests regarding quantum fluids is the description
of their dynamic properties. Ground state quantities are more
or less known as powerful tools such as the Monte Carlo or
the Variational methods have proven to succesfully describe
these properties. Now the challenge is to describe excitations.
Well to tell the truth muchinformation regarding elementary
excitations and collective modes in quantum liquids have been
gathered, and powerful methods based on variational models
have been succesfully applied to describe S(q,w) at low and
high momentum transfer. Still in quantum liquids are less
understood than groundtstae properties and there is room for
more work. The dynamic structure function is the maximum information
about the dynamics of the system one can get from an inelastic
neutron scattering experiment, so that is a good
starting point.

Path
Integral and Monte Carlo Simulations
Since the
beginning of 2006 I started to work on simulation methods at
the UPC. Monte Carlo and Path Integral monte Carlo are perhapshe
most prominent tools in the study of quantum systems right now.
Aside from statistical errors, these methods solve exactly
the manybody problem. Once again I'll be using them to study
mostly static properties of quantum fluids like Helium, BoseEinstein
condensates and/or weakly interacting Bose and fermi systems.
I am by no
means an expert on that, but that's going to be for the sure
the main tool I'll be using in my research on quantum fluids
for the next years. Right now that means learining and testing
algorithms, but taking into account the team of people I work
with, that's going to be a rewarding activity for sure.

Boltzmann
Machines
Before I
started to work at the Universitat politècnica de
catalunya I was an associate professor at Enginyeria
i Arquitectura La Salle from the Universitat Ramon Llull.
People there is much more interested in engineering that
in physics, so I decided to do some research in a filed that
could interest them and me. Neural networks was the perfect
field since many techniques commen in statistical mechanics
are used in the analysis of the computational capabilities of
neural networks. Boltzmann machines caught my interest due to
its unique property of learning probability distributions. What
began as sort of a game soon became a promising field of activity
where new learning techniques were developped and tested against
the performance of classical learning algorithms. 






