Abstract
We developed Convis, a Python simulation toolbox for large scale neural populations which offers arbitrary receptive fields by 3D convolutions executed on a graphics card. The resulting software proves to be flexible and easily extensible in Python, while building on the PyTorch library (The Pytorch Project, 2017), which was previously used successfully in deep learning applications, for just-in-time optimization and compilation of the model onto CPU or GPU architectures. An alternative implementation based on Theano (Theano Development Team, 2016) is also available, although not fully supported. Through automatic differentiation, any parameter of a specified model can be optimized to approach a desired output which is a significant improvement over e.g., Monte Carlo or particle optimizations without gradients. We show that a number of models including even complex non-linearities such as contrast gain control and spiking mechanisms can be implemented easily. We show in this paper that we can in particular recreate the simulation results of a popular retina simulation software VirtualRetina (Wohrer and Kornprobst, 2009), with the added benefit of providing (1) arbitrary linear filters instead of the product of Gaussian and exponential filters and (2) optimization routines utilizing the gradients of the model. We demonstrate the utility of 3d convolution filters with a simple direction selective filter. Also we show that it is possible to optimize the input for a certain goal, rather than the parameters, which can aid the design of experiments as well as closed-loop online stimulus generation. Yet, Convis is more than a retina simulator. For instance it can also predict the response of V1 orientation selective cells. Convis is open source under the GPL-3.0 license and available from https://github.com/jahuth/convis/ with documentation at https://jahuth.github.io/convis/.
Highlights
We developed Convis as an extension to the popular PyTorch toolbox (The Pytorch Project, 2017) that can implement models of responses to visual inputs on a large scale and, using automatic differentiation, derive the gradient of the output with respect to any input parameter
We show the utility of this method by examining models of retinal ganglion cells (RGCs), ranging from a very simple linear-non-linear model to a many-parameter mechanistic model which includes recurrent contrast gain control
Our implementation of the VirtualRetina model using Convis is available in the convis.retina submodule
Summary
We developed Convis as an extension to the popular PyTorch toolbox (The Pytorch Project, 2017) that can implement models of responses to visual inputs on a large scale and, using automatic differentiation, derive the gradient of the output with respect to any input parameter. Convis: Convolutional Vision Model Toolbox toolbox can simulate cell populations on a similar scale and in comparable time to previous large-scale retina simulators, even though the computations refrain from simplifying the shape of receptive fields. To understand the structures and processes of the visual system, e.g., how plasticity shapes representations (Masquelier, 2012), theorists are interested in a relatively large portion of the visual field, which requires a large number of neurons to be simulated according to our current understanding of the typical response of the respective cells. While the mechanisms of the model used are often similar, the theorist requires a large number of cells to be simulated rapidly and efficiently with identical parameters from the literature and the experimentalist needs a method to adjust model parameters such that the model recreates the behavior of the cells observed in experiments
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
