Marius Pachitariu
Gatsby Computational Neuroscience Unit
University College London
17 Queen Square
London WC1N 3AR, UK
Phone: +44 (0)20 7679 1192
Email: marius [at] gatsby [dot] ucl [dot] ac [dot] uk
About me
I am a PhD student at Gatsby, working with Maneesh Sahani
on probabilistic models of sequential data (video, text and spike trains)
based on recurrent neural networks.
I have previously completed a
mathematics undergraduate degree at Princeton University. My CV is here.
Research interests
- recurrent neural networks
- natural scene/video statistics
- statistical models of spike trains and decoding for BCI
- statistical language modelling
Talks
Learning visual motion in recurrent neural networks
Redwoood Center for Theoretical Neuroscience, Berkeley University
[video]
[slides]
Publications
Regularization and nonlinearities for neural language models: when are they needed?
Marius Pachitariu and Maneesh Sahani
arXiv:1301.5650 2013
[pdf]
Generalized recurrent linear models
Marius Pachitariu, Lars Buesing and Maneesh Sahani
Cosyne Abstracts 2013, Salt Lake City USA
[abstract]
Learning visual motion in recurrent neural networks
Marius Pachitariu and Maneesh Sahani
Advances in Neural Information Processing Systems 25 (NIPS 2012)
[pdf]
[poster]
Linear leaky integrate-and-fire model for gamma distributions of interspike intervals
Marius Pachitariu, Joseph K Jun, Jeffrey C Erlich, Bingni W Brunton and Carlos D Brody
Cosyne Abstracts 2011, Salt Lake City USA
[abstract]
[poster]
Spike train statistics
Marius Pachitariu
Princeton senior thesis
[pdf]