Git Product home page Git Product logo

ltts's Introduction

Learning Through Target Spikes (LTTS)

This is the Learning Through Target Spikes (LTTS) repository for code associated to the paper: Paolo Muratore, Cristiano Capone, Pier Stanislao Paolucci (2020) "Target spike patterns enable efficient and biologically plausible learning for complex temporal tasks" (published in PlosONE at: https://doi.org/10.1371/journal.pone.0247014).

Please give credit to this paper if you use or modify the code in a derivative work. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ or send a letter to Creative Commons, PO Box 1866, Mountain View, CA 94042, USA.

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

The Model

The model is implemented in the ltts.py module. A graphical depiction of the system architecture is offered in the Figure below.

Model

How to Run the Code

The code is written in Python 3.7 and for correct execution the program requires the following external dependences:

numpy
tqdm

To train the model (and save the results as Trained Model.npy) on the standard 3D trajectory task simply run:

python test.py

Results

We test our model on several tasks. See our article for full analysis and descriptions.

Temporal XOR

Here we report the results for the temporal XOR, which is a non-linear temporal task in which the network is asked to performed a delayed XOR computation given two inputs.

Temporal XOR

Walking Dynamics

We tested our system on a real-life task: the generation of a 56-dimensional walking dynamics. Our system was trained for 150ms and tested on 600ms, thus providing strong evidence of the generalization capabilities of this model.

Figure 4

Fast convergence

We tested our model on a smaller version of the 3D trajectory task to illustrate how our model achieves very fast convergence to good solutions. We also compare the performance of the online approximation to the exact gradient and found how it is beneficial for performances.

Figure 3

Robustness to Noise

To asses the robustness of our model to noise we corrupt the input current with a zero mean Gaussian noise of increasing variance. Indeed our voltage-dependent training rule proved to be rather stable under input perturbation.

Figure 2

ltts's People

Contributors

myscience avatar araihc avatar cristianocapone avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.