Luigi Acerbi

I am currently a 2nd year PhD candidate under the supervision of Sethu Vijayakumar (Edinburgh) and Daniel Wolpert (Cambridge). Before coming to Edinburgh, where I obtained a MSc in Neuroinformatics, I studied at the University of Milano-Bicocca (Milan, Italy), earning two BSc (Physics, Computer Science) and a MSc (Theoretical Physics). I am funded by the EPSRC and MRC under the Neuroinformatics DTC programme.

My broad research interest is the probabilistic structure of physical reality as reconstructed by an optimal inference machine - namely, the brain - and I am particularly interested in time (from physics to psychophysics).

My PhD project aims at characterizing the properties and dynamics of subjective sensorimotor space-time, combining probabilistic and computational modelling, machine learning techniques and psychophysical experiments. The guiding principle is that the structure of events in space-time is inferred, not given; while the probabilistic aspect of perception is becoming clear in many (spatial) domains, time is almost always treated as a known parameter, which is often a good approximation but it can break down under certain conditions.

The current work focuses on learning, adaptation and recalibration phenomena in sensorimotor timing tasks. The principal aim consists in clarifying the role of the statistical properties of the context (e.g. priors, likelihoods, loss functions, in Bayesian terms) in calibrating human sensorimotor timing, exploring within a theoretically sound framework unexplained phenomena such as temporal recalibration.

Keywords: Bayesian Brain, Machine Learning, Time and time perception, Subjective space-time, Psychophysics.

Contact

IPAB, School of Informatics, University of Edinburgh
Informatics Forum, 1.38
10 Crichton Street
Edinburgh EH8 9AB

E-mail: L.Acerbi [ANKH] sms.ed.ac.uk

Research Projects

• "Classical thermalization of scalar field theories and the reheating problem", MSc Theoretical Physics thesis, 2009.
  
• "Bayesian structural inference in sensorimotor temporal recalibration", MSc Neuroinformatics thesis, 2010.
• "Inferring the structure of space-time through action and perception", PhD Project.
  

Publications

• L. Acerbi and S. Vijayakumar, "Bayesian Causal Inference Drives Temporal Sensorimotor Recalibration", Proc. Computational and Systems Neuroscience COSYNE '11, Salt Lake City, Utah (2011).
• If you're interested in Cellular Automata, you might have a look here.