Sepp Kollmorgen

Position:
Visiting Scientist -- ended Dec 2022
Email:
Work phone:
Location:

Studying behavioral change using nearest neighbours

We developed a new method for analysing change in high-dimensional data based on nearest-neighbour statistics and applied it to song dynamics during vocal learning in zebra finches. The method could potentially be applied to other biological and artificial behaviours. (Paper)


Mixed pair HMMs for neural response modeling

Neural response models relate brain activity to peripheral signals such as motor outputs or sensory inputs. Mixed pair HMMs are a novel type of model that relates spiking behavior to peripheral signals in situations where the exact functional and temporal relationship between the signals evolves over time according to a Markov process with hidden states. (Paper) (Code)


Factor graph like models of the cortex

I worked on the hypothesis that the neo-cortex is implementing a factor graph like model and that its main function is to perform probabilistic inference. I studied a 'biologically plausible' self organizing neural network that organizes into a factor graph like structure when exposed to inputs. (More)

Supervisor

Valerio Mante

Publications

2022

2021

2020

2019

2015

  • Bhargava, S. and Blaettler, F. and Kollmorgen, S. and Liu, S.-C. and Hahnloser, R. H. Linear methods for efficient and fast separation of two sources recorded with a single microphone, Neural Computation, 22:(10), 2015

2014

2010