1,730,483 research outputs found
wenhao-z/Sampling_PoissSpk_Neuron: Sampling_Spike_Neuron
No full textSampling-based inference in a population of Poisson spiking neurons
Overview
User guide and system requirements
License
Overview
This code simulates a recurrent neural network and coupled networks of Poisson spiking neurons which implement sampling-based Bayesian inference to approximate the multivariate posterior of latent stimuli. In the network model, the sampling is driven by the internal Poisson variability of spike generation, and the structured recurrent connections store a prior of latent stimuli. We show the fixed (discrete) Poisson variability is able to sample different types of continuous distributions with different uncertainties.
The network has excitatory (E) and inhibitory (I) neurons, with each neuron modeled as a Hawkes process.
The E neurons have Gaussian tunings over the stimulus feature, while the I neurons don't have selectivity over stimulus and only keep the stability of the network.
Recurrent connections in the network
Two E neurons are connected according to their tuning similarity.
All other connections in the network, including E to I, I to E and I to I are unstructured.
More details of the network model and the results of this code can be found at our bioRxiv paper at:
https://www.biorxiv.org/content/10.1101/2020.11.18.389197v2
User guide and system Requirements
The code was developed on MATLAB R2018b and can be ran on MATLAB whose version is at least R2018b, no matter what operating system.
Directly download the whole package of code.
There are four main codes the user can use to reproduce the results in this study
simExpMdls.m
Cluster_code\simNet_ScanPars.m
Cluster_code\ansNetDat_ScanPars.m
Cluster_code\scanCoupledNet_RandPars_Cluster.m
Guideline of running simExpMdls.m
Directly run simExpMdls.m and you can get most of demo results in the presented study.
You can change the variable flagTask in simExpMdls.m to get the results under different demo tasks.
The simExpMdls.m can be ran on a standand computer with enough RAM. The computer should have at least 8GB of RAM.
Guideline of running Cluster_code\simNet_ScanPars.m
The simNet_ScanPars.m should be ran on a high performance cluster (HPC).
A single loop in simNet_ScanPars.m will take about 4GB of RAM and ran about 40 mins in a single thread.
The user may need to revise simNet_ScanPars.m based on the job scheduler or mangement on the HPC.
Guideline of running Cluster_code\ansNetDat_ScanPars.m
ansNetDat_ScanPars.m should be only ran after running Cluster_code\simNet_ScanPars.m to obtain a results.
You need to set fileName in ansNetDat_ScanPars.m to read the results ran by simNet_ScanPars.m.
License
This project is covered under the Apache 2.0 License
Wenhao-Jin/HydRA: v0.1.21.34
No full text<p>Fix the "UnboundLocalError: local variable 'G1' referenced before assignment" error when running train_and_evaluate_HydRa2_0.py with --no-PPI --no-PIA.</p>
CAPTCHA Recognition System
Get PDFCompletely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA), is a public fully automatic program that distinguishes users from computers or people. This thesis developed a CAPTCHA recognition system that can be deployed on the NAO robot, a humanoid robot to pass the Turing test and also can be deployed on web services in order to provide a recognition service.
The recognition system uses convolutional neural network to extract features in CAPTCHA image and encode data with one-hot encoding system which is widely used in multiclassification. Python is the programming language used in developing this project, TensorFlow and Keras library are used to easily establish a neural network. NAO robot version is v5 and code testing is on Ubuntu 16.04 release.
The final recognition model showed about 99.67% accuracy on train dataset and 98.10% accuracy on test dataset with suitable optimizer and loss function. According to the one-hot encoding features when regulated data, the accuracy is a bit high than it performed in real applications. Due to a large amount of CAPTCHA data for the combination of numbers and letters, the CAPTCHA in this thesis dataset consists only of numbers, which could be improved by using datasets contains numbers and letters CAPTCHA.Video Link: https://youtu.be/4pxHi9h1ca
[CODE] POPL24 Artifact for Soundly Handling Linearity
No full textThis artifact contains the implementation of the extension of Links with control-flow linearity as described in Section 4 of the paper: Wenhao Tang, Daniel Hillerström, Sam Lindley, J. Garrett Morris, "Soundly Handling Linearity", Proc. ACM Program. Lang. 8(POPL), 2024.Wenhao Tang. (2023). POPL24 Artifact for Soundly Handling Linearity (v1.0.0). Zenodo. https://doi.org/10.5281/zenodo.1012012
Trained model parameters for SynNet
No full textThis is a set of trained models for SynNet. Please see https://github.com/wenhao-gao/SynNet for more instruction
Results for practival molecular optimization (PMO) benchmark
No full textThis file contains a full set of results generated from practival molecular optimization (PMO) benchmark (https://github.com/wenhao-gao/mol_opt)</p
- …
