left-arrowleft-arrowright-arrowleft-arrowAsset 9

Sarah Marzen

Assistant Professor of Physics

Email: smarzen@kecksci.claremont.edu
Office: Keck Science Center 116
Phone: 909-607-3097
Web Site: https://sarahmarzen.weebly.com

Educational Background

B.S. in Physics, Caltech, 2011
Ph.D. in Physics, University of California, Berkeley, 2016

Research Interests

Biological organisms benefit from predicting their environmental input and have likely evolved sophisticated learning rules to do such prediction. At the same time, any organism’s ability to predict its input is limited by its resources: the size of its prediction apparatus, the energy it expends, and the capacity of the sensory channel that transduces information about past inputs.

How can we test how well biological organisms are predicting their input? And what are the learning rules by which such organisms improve their sensory prediction capabilities? A large part of my research is geared towards understanding how well and how organisms simultaneously predict and compress their sensory inputs. Such an understanding could lead to a better understanding of not just biological sensing, but also perhaps to new machine learning algorithms for prediction.

To date, I have investigated biologically appropriate metrics for the quality of a predictive biological sensor, establishing that sensory prediction can indeed improve fitness, and developed new methods for calculating limits to resource-limited prediction and for calculating the predictive and energetic performance of any given sensor. I have also spent considerable effort understanding perceptually relevant components of natural images and unfurling the implications of compressing large, random environments, towards the related goal of understanding compression sans prediction.

Selected Publications

  1. S. Marzen and J.P. Crutchfield. (2018). Optimized bacteria are environmental prediction engines. Physical Review E  98.
  2. S. Marzen and J.P. Crutchfield. (2017). Informational and causal architecture of continuous-time renewal processes. Journal of Statistical Physics 168: 109-127.
  3. S. Marzen and J.P. Crutchfield. (2017). Structure and randomness of continuous-time discrete-event processes. Journal of Statistical Physics 169: 303-315.
  4. S. Marzen and S. DeDeo. (2017). The evolution of lossy compression. Journal of the Royal Society Interface 14.
  5. S. Marzen. (2017). Difference between memory and prediction in linear recurrent networks. Physical Review E 96.
  6. S. Marzen and J.P. Crutchfield. (2017). Nearly maximally predictive features and their dimensions. Physical Review E (R)  95 .
  7. C. Hillar and S. Marzen. (2017). Neural network coding of natural images with applications to pure mathematics. Proceedings of the AMS Special Session on Algebraic and Geometric Methods in Discrete Mathematics Heather Harrington, Mohamed Omar, and Matthew Wright.
  8. C. Hillar and S. Marzen. (2016). Revisiting perceptual distortion for natural images: mean discrete structural similarity index. Data Compression Conference.
  9. S. Marzen and J.P. Crutchfield. (2016). Predictive rate-distortion for infinite-order Markov processes. Journal of Statistical Physics 163: 1312-1338.
  10. S. Marzen and S. DeDeo. (2016). Weak universality in sensory tradeoffs. Physical Review E (R)  94.
  11. S. Marzen and J.P. Crutchfield. (2016). Statistical Signatures of Structural Organization: The case of long memory in renewal processes. Physics Letters A   380: 1517-1525.
  12. S. Marzen and J.P. Crutchfield. (2015). Informational and causal architecture of discrete-time renewal processes. Entropy 17: 4891-4917.
  13. S. Marzen, M.R. DeWeese, and J.P. Crutchfield. (2015). Time resolution dependence of spike train information measures. Frontiers in Computational Neuroscience 9.
  14. J. P. Crutchfield and S. Marzen. (2015). Signatures of Infinity: Nonergodicity in Prediction, Complexity, and Learning. Physical Review E (R) 91.
  15. S. Marzen and J.P. Crutchfield. (2014). Information anatomy of stochastic equilibria. Entropy 16: 4713-4748.
  16. S. Marzen, H.G. Garcia, and R. P. Phillips. (2013). Statistical Mechanics of the Monod-Wyman-Changeux (MWC) Models. Journal of Molecular Biology 425: 1433-1460.
    Abstract – The 50th anniversary of the classic Monod-Wyman-Changeux (MWC) model provides an opportunity to survey the broader conceptual and quantitative implications of this quintessential biophysical model. With the use of statistical mechanics, the mathematical implementation of the MWC concept links problems that seem otherwise to have no ostensible biological connection including ligand-receptor binding, ligand-gated ion channels, chemotaxis, chromatin structure and gene regulation. Hence, a thorough mathematical analysis of the MWC model can illuminate the performance limits of a number of unrelated biological systems in one stroke. The goal of our review is twofold. First, we describe in detail the general physical principles that are used to derive the activity of MWC molecules as a function of their regulatory ligands. Second, we illustrate the power of ideas from information theory and dynamical systems for quantifying how well the output of MWC molecules tracks their sensory input, giving a sense of the “design” constraints faced by these receptors.