Contact

Sebastian Gerwinn

Adresse: Spemannstr. 41
72076 Tübingen
Raum Nummer: 1.B.05

 

Bild von Gerwinn, Sebastian

Sebastian Gerwinn

Position: Doktorand  Abteilung:  Alumni Bethge

My research interests are in the area of Bayesian inference and computational neuroscience. In particular, I am interested in a characterization of the relationship between sensory signals and neural responses. The methods which I think are the most promising to tackle this kind of tasks are Bayesian methods.

The applicability of Bayesian methods is often limited by the fact that they are computational prohibitive. My main focus has therefore been to alleviate this problem by developing approximate methods which are then also feasible on a much larger scale and can therefore be applied to realistically sized data.

The main advantage of a Bayesian treatment lies in the explicit representation of the involved uncertainties. Having access to this kind of knowledge enables one to perform further analysis such as experimental design or model selection.

 

 

Stimulus Response Relationship

I have analyzed the relationship between stimuli and neural responses from three different perspectives: (1) the encoding, (2) the decoding and (3) the joint occurrence perspective.


In a first project I investigated the system identification task corresponding to the encoding direction of the stimulus response relationship. I developed an approximate Bayesian inference method which is feasible for models of generalized linear type, one of the most successful and commonly used generative models. As a result, we obtained not only particular point estimates of sets of parameters, but also model based confidence intervals, which in turn we used for feature selection and estimating the functional connectivity within populations of neurons.


Second, I analyzed the relationship from a decoding point of view. Here, using the leaky integrate and fire neuron model, I obtained a simple yet accurate decoding algorithm. Again, using a Bayesian treatment, it is possible to not only decode the most likely stimulus but also assigning to each stimulus the probability that it has caused the observed neural response.


Third, merging both perspectives, I looked at the joint occurrence of stimuli and neural responses. Using commonly used descriptive statistics such as spike-triggered average and spike-triggered covariance, I build a maximum entropy model. This model can then be used as a generative model as well as a decoding model exhibiting the same descriptive statistics as the observed ones, while assuming the least  additional constrains due to the maximum entropy property.

Unüberwachtes Lernen Steuerbarer Filter
Unüberwachtes Lernen Steuerbarer Filter
Obwohl sich die Pixel-Darstellung eines Bildes unter affinen Transformationen wie Translation,     Rotation und Skalierung stark ändert, bleibt der Inhalt des Bildes weitgehend unverändert. Insbesondere, wenn sich die Änderungen eines D-dimensionalen Lichtintensitätsvektors durch eine einparametrige Lie-Gruppe beschreiben lassen, ist es möglich, eine verlustfreie Bildrepräsentation zu finden, bei der eine Komponente dem Transformationsparameter entspricht und die anderen (D-1) Komponenten invariant sind unter der Lie-Gruppentransformation. Um solche Bildrepräsentationen abzuleiten, konstruieren wir geeignete generative Modelle, mit denen Steuerbare Filter auf unüberwachte Weise gelernt werden können. Insbesondere haben wir zeigen können, dass es möglich ist mit einer anti-symmetrischen Variante der Kanonischen Korrelationsanalyse (CCA), eine vollständige Basis für 32x32 Bildausschnitte zu bestimmen, die sich aus rotationsinvarianten Steuerbaren Filtern zusammensetzt.

Bayesian Models for Multi-Electrode Neuronal Spike Recordings

We investigate Bayesian methods to predict responses from multiple retinal and LGN ganglion cells, conditioned on visual stimuli.

Präferenzen: 
Referenzen pro Seite: Jahr: Medium:

  
Zeige Zusammenfassung

Artikel (9):

Sinz FH, Lies J-P, Gerwinn S und Bethge M (November-2014) Natter: A Python Natural Image Statistics Toolbox Journal of Statistical Software 61(5) 1-34.
Haefner RM, Gerwinn S, Macke JH und Bethge M (Februar-2013) Inferring decoding strategies from choice probabilities in the presence of correlated variability Nature Neuroscience 16(2) 235–242.
Theis L, Gerwinn S, Sinz F und Bethge M (November-2011) In All Likelihood, Deep Belief Is Not Enough Journal of Machine Learning Research 12 3071-3096.
Macke JH, Gerwinn S, White LW, Kaschube M und Bethge M (Mai-2011) Gaussian process methods for estimating cortical maps NeuroImage 56(2) 570-581.
Berens P, Ecker AS, Gerwinn S, Tolias AS und Bethge M (März-2011) Reassessing optimal neural population codes with neurometric functions Proceedings of the National Academy of Sciences of the United States of America 108(11) 4423-4428.
Gerwinn S, Macke JH und Bethge M (Februar-2011) Reconstructing stimuli from the spike-times of leaky integrate and fire neurons Frontiers in Neuroscience 5(1) 1-16.
Gerwinn S, Macke J und Bethge M (April-2010) Bayesian inference for generalized linear models for spiking neurons Frontiers in Computational Neuroscience 4(12) 1-17.
Gerwinn S, Macke JH und Bethge M (Oktober-2009) Bayesian population decoding of spiking neurons Frontiers in Computational Neuroscience 3(21) 1-14.
Sinz FH, Gerwinn S und Bethge M (Mai-2009) Characterization of the p-Generalized Normal Distribution Journal of Multivariate Analysis 100(5) 817-820.

Beiträge zu Tagungsbänden (6):

Gerwinn S, Berens P und Bethge M (April-2010) A joint maximum-entropy model for binary neural population patterns and continuous signals In: Advances in Neural Information Processing Systems 22, , 23rd Annual Conference on Neural Information Processing Systems (NIPS 2009), Curran, Red Hook, NY, USA, 620-628.
pdf
Macke JH, Gerwinn S, Kaschube M, White LE und Bethge M (April-2010) Bayesian estimation of orientation preference maps In: Advances in Neural Information Processing Systems 22, , 23rd Annual Conference on Neural Information Processing Systems (NIPS 2009), Curran, Red Hook, NY, USA, 1195-1203.
pdf
Berens P, Gerwinn S, Ecker AS und Bethge M (April-2010) Neurometric function analysis of population codes In: Advances in Neural Information Processing Systems 22, , 23rd Annual Conference on Neural Information Processing Systems (NIPS 2009), Curran, Red Hook, NY, USA, 90-98.
pdf
Gerwinn S, Macke J, Seeger M und Bethge M (September-2008) Bayesian Inference for Spiking Neuron Models with a Sparsity Prior In: Advances in neural information processing systems 20, , Twenty-First Annual Conference on Neural Information Processing Systems (NIPS 2007), Curran, Red Hook, NY, USA, 529-536.
pdf
Seeger M, Gerwinn S und Bethge M (September-2007) Bayesian Inference for Sparse Generalized Linear Models In: Machine Learning: ECML 2007, , 18th European Conference on Machine Learning, Springer, Berlin, Germany, 298-309, Series: Lecture Notes in Computer Science ; 4701.
Bethge M, Gerwinn S und Macke JH (Februar-2007) Unsupervised learning of a steerable basis for invariant image representations In: Human Vision and Electronic Imaging XII, , SPIE Human Vision and Electronic Imaging Conference 2007, SPIE, Bellingham, WA, USA, 1-12, Series: Proceedings of the SPIE ; 6492.
pdf

Poster (10):

Haefner RM, Gerwinn S, Macke JH und Bethge M (November-2011): Relationship between decoding strategy, choice probabilities and neural correlations in perceptual decision-making task, 41st Annual Meeting of the Society for Neuroscience (Neuroscience 2011), Washington, DC, USA.
Berens P, Ecker AS, Gerwinn S, Tolias AS und Bethge M (Februar-2011): Optimal Population Coding, Revisited, Computational and Systems Neuroscience Meeting (COSYNE 2011), Salt Lake City, UT, USA.
Theis L, Gerwinn S, Sinz F und Bethge M (Oktober-2010): Likelihood Estimation in Deep Belief Networks, Bernstein Conference on Computational Neuroscience (BCCN 2010), Berlin, Germany, Frontiers in Computational Neuroscience, 2010(Conference Abstract: Bernstein Conference on Computational Neuroscience).
Häfner R, Gerwinn S, Macke JH und Bethge M (Oktober-1-2009): Neuronal decision-making with realistic spiking models, Bernstein Conference on Computational Neuroscience (BCCN 2009), Frankfurt a.M., Germany, Frontiers in Computational Neuroscience, 2009(Conference Abstract: Bernstein Conference on Computational Neuroscience) 132-133.
Macke J, Gerwinn S, White L, Kaschube M und Bethge M (März-2009): Bayesian estimation of orientation preference maps, Computational and Systems Neuroscience Meeting (COSYNE 2009), Salt Lake City, UT, USA, Frontiers in Systems Neuroscience, 2009(Conference Abstracts: Computational and Systems Neuroscience).
Gerwinn S, Macke J und Bethge M (März-2009): Bayesian Population Decoding of Spiking Neurons, Computational and Systems Neuroscience Meeting (COSYNE 2009), Salt Lake City, UT, USA,, Frontiers in Systems Neuroscience, 2009(Conference Abstracts: Computational and Systems Neuroscience).
Gerwinn S (Juli-31-2008): Bayesian decoding of populations of integrate-and-fire neurons, Gordon Research Conference: Sensory Coding & The Natural Environment 2008, Lucca, Italy.
Gerwinn S, Seeger M, Zeck G und Bethge M (April-2007): Bayesian Neural System identification: error bars, receptive fields and neural couplings, 7th Meeting of the German Neuroscience Society, 31st Göttingen Neurobiology Conference, Göttingen, Germany, Neuroforum, 13(Supplement) 360.
pdf
Bethge M, Macke JH, Gerwinn S und Zeck G (April-2007): Identifying temporal population codes in the retina using canonical correlation analysis, 7th Meeting of the German Neuroscience Society, 31st Göttingen Neurobiology Conference, Göttingen, Germany, Neuroforum, 13(Supplement) 359.
pdf
Gerwinn S, Seeger M, Zeck G und Bethge M (Februar-2007): Bayesian Receptive Fields and Neural Couplings with Sparsity Prior and Error Bars, Computational and Systems Neuroscience Meeting (COSYNE 2007), Salt Lake City, UT, USA.

Vorträge (4):

Gerwinn S, Macke JH und Bethge M (Oktober-2010) Abstract Talk: Toolbox for inference in generalized linear models of spiking neurons, Bernstein Conference on Computational Neuroscience (BCCN 2010), Berlin, Germany, Frontiers in Computational Neuroscience, 2010(Conference Abstract: Bernstein Conference on Computational Neuroscience).
Haefner R, Gerwinn S, Macke J und Bethge M (Februar-2010) Abstract Talk: Implications of correlated neuronal noise in decision making circuits for physiology and behavior, Computational and Systems Neuroscience Meeting (COSYNE 2010), Salt Lake City, UT, USA, Frontiers in Neuroscience, Conference Abstract: Computational and Systems Neuroscience 2010.
Berens P, Gerwinn S, Ecker AS und Bethge M (Oktober-1-2009) Abstract Talk: Neurometric function analysis of short-term population codes, Bernstein Conference on Computational Neuroscience (BCCN 2009), Frankfurt a.M., Germany, Frontiers in Computational Neuroscience, 2009(Conference Abstract: Bernstein Conference on Computational Neuroscience) 24-25.
Gerwinn S, Seeger M, Zeck G und Bethge M (November-2006) Abstract Talk: Bayesian Neural System identification: error bars, receptive fields and neural couplings, 7th Conference of the Junior Neuroscientists of Tübingen (NeNa 2006), Oberjoch, Germany 9.

Export als:
BibTeX, XML, pubman, Edoc, RTF
Last updated: Montag, 22.05.2017