Bridging biophysics and computation with differentiable simulation
Michael Deistler, Kyra Kadhim, Jonas Beck, Matthijs Pals, Janne Lappalainen, Manuel Gloeckler, Ziwei Huang, Cornelius Schroeder, Philipp Berens, Pedro Gonçalves, Jakob Macke · Bernstein Conference 2024
Biophysical neuron models provide mechanistic insight about empirically observed phenomena. However, biophysical neuron models are expensive to simulate, thereby limiting the scale of biophysical models, and optimizing their parameters is notoriously difficult, thereby preventing the fitting of biophysical models to physiologically meaningful tasks or datasets (panel a). We built a new toolbox for simulation and inference in neuroscience, Jaxley, which allows to parallelize simulations on (multiple) GPUs and to compute the gradient of simulations with respect to parameters (panel b, c). We use