My current research focuses on trying to understand biological systems as solutions to computational problems. When I’m not doing research, I often think about how to improve science and education through the web.
More on my research interests.
Current Projects
Research
Modeling dendritic integration and somatic membrane potential through hierarchical GLMs
I’m working with Máté Lengyel at Cambridge to build a simple yet flexible, canonical model for how the dendrites of neurons integrate their inputs. The model is based on a hierarchy of generalized linear models (GLMs), which have recently been successfully applied to describe the spiking responses of sensory neurons. We are testing the appropriateness of this approach by fitting data from simulated neurons and 2-photon glutamate uncaging experiments. The hope is to use these models to better understand computation in single neurons and small populations of neurons, including whether and how one neuron might infer (statistics of) the analog membrane potential of upstream neurons using only discrete action potentials.
More on dendritic integration and GLMs.
Sniff-Modulations of the Olfactory Bulb Vocabulary
I’m working with Elad Schneidman, Jakob Macke, and Dima Rinberg to understand how neural activity in the olfactory bulb of mice represents information about their olfactory environment and behavior. In particular, we are studying how sniffing modifies the typical patterns of neural activity and what this might tell us about the links between action and perception in olfaction. This collaboration began as my student project at the Methods in Computational Neuroscience course at the Marine Biological Laboratory in Woods Hole, MA.
Optimal Dendrite Morphology for One-Shot Learning Tasks
I’m working with Professor Bartlett Mel and graduate student Xundong Wu at USC to understand how brains rapidly and robustly encode information presented only once. In particular, we’re investigating the optimal dendrite morphology for memory capacity during one-shot learning tasks and studying how the optimal morphology varies with input features such as noise and density of activation.
More on dendrite morphology and memory capacity.
Past Projects
Research
Reliable Brains from Unreliable Neurons – The Search for Synfire Chains in the Brain
I worked with Professor Kwabena Boahen’s group at Stanford to study the emergence of synfire chains in biological neural networks. This work was sponsored by the Stanford Amgen Scholars Program and I presented my findings at the Stanford Amgen Scholars Symposium.
More on reliable brains from unreliable neurons.
The Quest for Real-Time Simulations of Neural Synapses
I helped USC Professor Ted Berger‘s synaptic modeling group speed up their simulations by borrowing numerical integration techniques from computational physics. Our work paved the way for an ongoing series of joint projects between the Berger group and the computational condensed matter group in the physics department, the first such link between the neuroscience and physics communities at USC.
A Levinson’s Theorem for Scattering on Graphs
I worked with Professor Andrew Childs from the Institute for Quantum Computing and University of Waterloo in Ontario, Canada to extend the continuous Levinson’s theorem for scattering to a discrete version for scattering on graphs. The theorem relates the number of bound states of a potential to the winding of the phase shift of the scattering states and (we hope) may provide inspiration for new algorithms for quantum computers.
The Nature of Equilibration in the Quantum World
I worked with USC Professor Paolo Zanardi and the Quantum Information Group at the Institute for Scientific Interchange in Torino, Italy to define and justify an appropriate concept of equilibration in quantum mechanics. The unitary dynamics of quantum mechanics do not allow convergence in a strong sense of quantum states to equilibrium states so alternative definitions are required. This work was sponsored by a USC Provost Summer Research Fellowship.
More on quantum statistical mechanics.
Web Development
CoLab
Fed up with the lack of good web tools for scientific collaboration, Casey Stark and I cooked up a platform for scientific collaboration on the web called CoLab. The ultimate goal is to enable online collaboration around any piece of scientific content and to increase the speed of information-sharing among scientists and between scientists and the public. We debuted CoLab at the Open Science Summit 2010 in Berkeley and are currently working to add additional features. The project is open source and we’re always looking for collaborators.
College: What I Did Right and Where I Screwed Up // Jun 1, 2011 at 6:03 pm
[...] projects [...]