Zachary Kilpatrick Surges in Productivity During Pandemic Times
Read on to be inspired by this ICS Fellow.
Public Dissemination
The article in Physical Review Letters focuses on how diversity in networks of decision making agents can increase decision efficiency via more deliberate agent’s observation of the split of hasty observers’ decisions. This work is of relevance to the propagation of social information on social media, especially for binary decisions like voting in two-party systems. It was recently featured both in an article on CU Boulder Today and on the 6pm Denver 7 TV news.
Student Publications
- Postdoc Tahra Eissa is completing a paper focusing on biases in humans predicting rare evidence for which she leveraged a combination of Bayesian modeling, model identification, model-free analysis, and data from over 200 human subjects collected on Amazon MTurk. We will submit this work soon to Nature Human Behavior.
- 4th year PhD student Subekshya Bidari is completing a paper on the effect of hive geometry on the dynamics of collective foraging decisions in honeybee colonies. This work will be submitted soon to the Journal of Mathematical Biology. Subekshya was also recently awarded a four month DAAD: German Academic Exchange Service Dissertation Fellowship to work with collaborators at the University of Konstanz/Max Planck Institute for Animal Behavior using animal tracking data from primates and bats to constrain mechanistic models of social foraging behavior. She plans to go from April to August 2021.
- 3rd year PhD student Nick Barendregt is completing a paper on model-based analysis of human use of adaptive decision making rules in tasks with dynamic reward and evidence quality conditions. He has fit his models to data from a prediction task in which humans must guess which of two buckets will fill up with the most tokens based on an observed history of token movements. He is also helping to develop future experiments to test further model predictions with our collaborator Josh Gold at University of Pennsylvania.
- 2nd year PhD student Heather Cihak has started on a project focusing on the role of short term plasticity in shaping working memory degradation in large-scale neural circuit models of persistent activity.
- Masters student Tim Thorn will defend his thesis, "Learning algorithms for biologically plausible recurrent neural networks” on November 17. His work focuses on uncovering dynamical systems principles for how machine learning methods train recurrent neural networks to perform cognitive tasks.
- Former PhD student Daniel Poll began a tenure track position at College of Charleston in the Department of Mathematics.
Collaboration
- I have been working in collaboration with Stefan Mihalas at the Allen Institute on a project exploring the role of activity frequency in local field potentials in shaping the functional connectivity of mouse visual cortex. We are leveraging data from 30 subjects recorded using high throughput Neuropixels in six different visual areas.
- I have also been collaborating with Abigail Person at CU Anschutz to study how information propagates through feedforward architecture in the cerebellum from mossy fibers to granule cells, and how the combination of input summation and thresholding can lead to both good sparsification for target function learning while still preserving input information.
Presentations & Talks
- “Heterogeneity improves speed and accuracy in social networks” in Northwestern University, Engineering Sciences & Applied Mathematics Colloquium, virtual, 11/2020
- “Normative theory of urgency in environments with dynamic context” at Bernstein Computational Neuroscience Conference: Workshop on dynamic probabilistic inference in the brain, virtual, 9/2020
- “Patch leaving decisions as a first exit time problem” in Brandeis University, Mathematical Biology Seminar, virtual, 6/2020
- “Normative theory of patch foraging decisions” in Baylor College of Medicine/Rice University, Theoretical Neuroscience Seminar, virtual, 5/2020
Conference Organization
- Math + Neuroscience: Strengthening the interplay between theory and mathematics, (co-organizer; main organizers: Carina Curto and Katie Morrison) Semester-Long Program at the Institute for Computational and Experimental Research in Mathematics, Brown University, Providence, Rhode Island [proposed for Fall 2023]
- Dynamical principles of biological and artificial neural networks, (with Sue Ann Campbell, Alona Fyshe, and Joel Zylberberg) Five Day Workshop at the Banff International Research Station (≈ 40 participants), Banff, Alberta, Canada [proposed for 2022]
- International Conference on Mathematical Neuroscience,
- Advisory Committee, Leipzig, Germany, 7/2021
- Advisory Committee (34 invited; 970 attendees), Virtual (Hosted on Zoom/Youtube), 6/2020