Our paper: "Statistical analysis and optimality of neural systems" is now out in Neuron
New preprint: "Attentional dynamics of efficient codes" is available on bioRxiv.
We are organizing the Inferring and testing optimality in perception and neurons workshop at the Bernstein Conference 2020. The event will be held online on Sept 29th 2020.
"Top-down attention as efficient perceptual inference"
with Gašper Tkačik
"Hierarchical inference guides perceptual decision-making
in a dynamic environment" with Julie Charlton, Yoon Bai,
Ann Hermundstad and Robbe Goris
"Co-occurrence statistics of natural sound features
predict perceptual grouping" with Josh McDermott
"Sensory codes for optimizing tradeoffs between task
performance, adaptation speed and resource use"
with Ann Hermundstad
Our paper: "Learning mid-level auditory codes from natural sound statistics" is out in Neural Computation
I'm organizing the Natural Scene Statistics and Sensory Representations workshop at the Bernstein Conference in Berlin on 21st Sept 2016
Education and working experience:
2018 - to date, IST Plus Fellow, Institute of Science and Technology Austria, Klosterneuburg, Austria
2015 - 2018, Postdoctoral Associate, Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, USA
2011 - 2015, PhD in Computer Science, Max-Planck Institute for Mathematics in the Sciences, Leipzig, Germany
2009 - 2011, Research Assistant in Bioinformatics, Department of Molecular Neuropharmacology, Institute of Pharmacology, Polish Academy of Sciences, Kraków, Poland
2005 - 2010 MSc in Computer Science, Jagiellonian University, Kraków, Poland
An important property separating living systems from inorganic matter is the ability to build and maintain internal models of the world. In order to achieve that, organisms extract regularities present in environments in which they evolved and developed.
The brain seems to be a prominent example of a system employing such a strategy. It has been demonstrated that numerous properties of perception and sensation can be explained as an adaptation to the natural environment.
In my work I follow these lines of thought. I analyze the statistical structure of natural stimuli, develop optimal processing strategies which might be approximated by biological systems and verify theoretical predictions through experimental data analysis. I hope that this approach will bring us towards identifying general principles which govern information processing in Nature.
Preprints and manuscripts under review:
Mlynarski W., Tkacik G., "Attentional dynamics of efficient codes", bioRxiv preprint, (preprint link)
Mlynarski W.*, Hermundstad A.*, "Efficient and adaptive sensory codes", Nature Neuroscience - in press, (preprint link)
Mlynarski W., McDermott J.H., "Ecological origins of perceptual grouping principles in the auditory system", Proceedings of the National Academy of Sciences, 2019 (link)
Mlynarski W., McDermott J.H., "Learning mid-level auditory codes from natural sound statistics", Neural Computation, 2018 (link)
Mlynarski W. "The opponent channel population code of sound location is an efficient representation of natural stereo sounds", PLOS Computational Biology, 2015 (link)
Mlynarski W., Jost J. "Statistics of natural binaural sounds", PLOS One, 2014 (link)
Mlynarski W, "Efficient coding of spectrotemporal binaural sounds leads to emergence of the auditory space representation", Frontiers in Computational Neuroscience, 2014 (link)
Mlynarski W., Freigang C., Bennemann J., Stoehr M. and Ruebsamen R.. "Position of acoustic stimulus modulates visual alpha activity", NeuroReport, 2014 (link)
Korostynski M., Piechota M., Dzbek J., Mlynarski W., Szklarczyk K., Ziolkowska B. and Przewlocki R. "Novel drug-regulated transcriptional networks in brain reveal pharmacological properties of psychotropic drugs", BMC Genomics, 2013 (link)
Conference Papers and Technical Reports:
Mlynarski W., McDermott J. H. "Natural sound statistics predict auditory grouping principles", CCN 2018, Philadelphia, USA, Proceedings
Mlynarski W., "Sparse, complex-valued representations of natural sounds learned with phase and amplitude continuity priors", arXiv:1312.4695 [cs.LG]
Mlynarski W., "Learning binaural spectrogram features for azimuthal speaker localization", Interspeech 2013, Lyon, France