Code repository of the paper Neural circuit policies enabling auditable autonomy published in Nature Machine Intelligence

Overview

Neural Circuit Policies Enabling Auditable Autonomy

DOI

Online access via SharedIt

Neural Circuit Policies (NCPs) are designed sparse recurrent neural networks based on the LTC neuron and synapse model loosely inspired by the nervous system of the organism C. elegans. This page is a description of the Keras (TensorFlow 2 package) reference implementation of NCPs. For reproducibility materials of the paper see the corresponding subpage.

alt

Installation

Requirements:

  • Python 3.6
  • TensorFlow 2.4
  • (Optional) PyTorch 1.7
pip install keras-ncp

Update January 2021: Experimental PyTorch support added

With keras-ncp version 2.0 experimental PyTorch support is added. There is an example on how to use the PyTorch binding in the examples folder and a Colab notebook linked below. Note that the support is currently experimental, which means that it currently misses some functionality (e.g., no plotting, no irregularly sampled time-series,etc. ) and might be subject to breaking API changes in future updates.

Breaking API changes between 1.x and 2.x

The TensorFlow bindings have been moved to the tf submodule. Thus the only breaking change regarding the TensorFlow/Keras bindings concern the import

# Import shared modules for wirings, datasets,...
import kerasncp as kncp
# Import framework-specific binding
from kerasncp.tf import LTCCell      # Use TensorFlow binding
(from kerasncp.torch import LTCCell  # Use PyTorch binding)

Colab notebooks

We have created a few Google Colab notebooks for an interactive introduction to the package

Usage: the basics

The package is composed of two main parts:

  • The LTC model as a tf.keras.layers.Layer or torch.nn.Module RNN cell
  • An wiring architecture for the LTC cell above

The wiring could be fully-connected (all-to-all) or sparsely designed using the NCP principles introduced in the paper. As the LTC model is expressed in the form of a system of ordinary differential equations in time, any instance of it is inherently a recurrent neural network (RNN).

Let's create a LTC network consisting of 8 fully-connected neurons that receive a time-series of 2 input features as input. Moreover, we define that 1 of the 8 neurons acts as the output (=motor neuron):

from tensorflow import keras
import kerasncp as kncp
from kerasncp.tf import LTCCell

wiring = kncp.wirings.FullyConnected(8, 1)  # 8 units, 1 motor neuron
ltc_cell = LTCCell(wiring) # Create LTC model

model = keras.Sequential(
    [
        keras.layers.InputLayer(input_shape=(None, 2)), # 2 input features
        keras.layers.RNN(ltc_cell, return_sequences=True),
    ]
)
model.compile(
    optimizer=keras.optimizers.Adam(0.01), loss='mean_squared_error'
)

We can then fit this model to a generated sine wave, as outlined in the tutorials (open in Google Colab).

alt

More complex architectures

We can also create some more complex NCP wiring architecture. Simply put, an NCP is a 4-layer design vaguely inspired by the wiring of the C. elegans worm. The four layers are sensory, inter, command, and motor layer, which are sparsely connected in a feed-forward fashion. On top of that, the command layer realizes some recurrent connections. As their names already indicate, the sensory represents the input and the motor layer the output of the network.

We can also customize some of the parameter initialization ranges, although the default values should work fine for most cases.

ncp_wiring = kncp.wirings.NCP(
    inter_neurons=20,  # Number of inter neurons
    command_neurons=10,  # Number of command neurons
    motor_neurons=5,  # Number of motor neurons
    sensory_fanout=4,  # How many outgoing synapses has each sensory neuron
    inter_fanout=5,  # How many outgoing synapses has each inter neuron
    recurrent_command_synapses=6,  # Now many recurrent synapses are in the
    # command neuron layer
    motor_fanin=4,  # How many incoming synapses has each motor neuron
)
ncp_cell = LTCCell(
    ncp_wiring,
    initialization_ranges={
        # Overwrite some of the initialization ranges
        "w": (0.2, 2.0),
    },
)

We can then combine the NCP cell with arbitrary keras.layers, for instance to build a powerful image sequence classifier:

height, width, channels = (78, 200, 3)

model = keras.models.Sequential(
    [
        keras.layers.InputLayer(input_shape=(None, height, width, channels)),
        keras.layers.TimeDistributed(
            keras.layers.Conv2D(32, (5, 5), activation="relu")
        ),
        keras.layers.TimeDistributed(keras.layers.MaxPool2D()),
        keras.layers.TimeDistributed(
            keras.layers.Conv2D(64, (5, 5), activation="relu")
        ),
        keras.layers.TimeDistributed(keras.layers.MaxPool2D()),
        keras.layers.TimeDistributed(keras.layers.Flatten()),
        keras.layers.TimeDistributed(keras.layers.Dense(32, activation="relu")),
        keras.layers.RNN(ncp_cell, return_sequences=True),
        keras.layers.TimeDistributed(keras.layers.Activation("softmax")),
    ]
)
model.compile(
    optimizer=keras.optimizers.Adam(0.01),
    loss='sparse_categorical_crossentropy',
)
@article{lechner2020neural,
  title={Neural circuit policies enabling auditable autonomy},
  author={Lechner, Mathias and Hasani, Ramin and Amini, Alexander and Henzinger, Thomas A and Rus, Daniela and Grosu, Radu},
  journal={Nature Machine Intelligence},
  volume={2},
  number={10},
  pages={642--652},
  year={2020},
  publisher={Nature Publishing Group}
}
You might also like...
Code to use Augmented Shapiro Wilks Stopping, as well as code for the paper "Statistically Signifigant Stopping of Neural Network Training"

This codebase is being actively maintained, please create and issue if you have issues using it Basics All data files are included under losses and ea

This repository contains the code for EMNLP-2021 paper "Word-Level Coreference Resolution"

Word-Level Coreference Resolution This is a repository with the code to reproduce the experiments described in the paper of the same name, which was a

Code for our paper
Code for our paper "Mask-Align: Self-Supervised Neural Word Alignment" in ACL 2021

Mask-Align: Self-Supervised Neural Word Alignment This is the implementation of our work Mask-Align: Self-Supervised Neural Word Alignment. @inproceed

Code for the paper: Sequence-to-Sequence Learning with Latent Neural Grammars

Code for the paper: Sequence-to-Sequence Learning with Latent Neural Grammars

Code to reprudece NeurIPS paper: Accelerated Sparse Neural Training: A Provable and Efficient Method to Find N:M Transposable Masks

Accelerated Sparse Neural Training: A Provable and Efficient Method to FindN:M Transposable Masks Recently, researchers proposed pruning deep neural n

Easy to use, state-of-the-art Neural Machine Translation for 100+ languages

EasyNMT - Easy to use, state-of-the-art Neural Machine Translation This package provides easy to use, state-of-the-art machine translation for more th

Open Source Neural Machine Translation in PyTorch
Open Source Neural Machine Translation in PyTorch

OpenNMT-py: Open-Source Neural Machine Translation OpenNMT-py is the PyTorch version of the OpenNMT project, an open-source (MIT) neural machine trans

Sequence-to-sequence framework with a focus on Neural Machine Translation based on Apache MXNet

Sockeye This package contains the Sockeye project, an open-source sequence-to-sequence framework for Neural Machine Translation based on Apache MXNet

Releases(v2.0.0)
Owner
PhD candidate at IST Austria. Working on Machine Learning, Robotics, and Verification
Problem: Given a nepali news find the category of the news

Classification of category of nepali news catorgory using different algorithms Problem: Multiclass Classification Approaches: TFIDF for vectorization

pudasainishushant 2 Jan 09, 2022
AllenNLP integration for Shiba: Japanese CANINE model

Allennlp Integration for Shiba allennlp-shiab-model is a Python library that provides AllenNLP integration for shiba-model. SHIBA is an approximate re

Shunsuke KITADA 12 Feb 16, 2022
【原神】自动演奏风物之诗琴的程序

疯物之诗琴 读取midi并自动演奏原神风物之诗琴。 可以自定义配置文件自动调整音符来适配风物之诗琴。 (原神1.4直播那天就开始做了!到现在才能放出来。。) 如何使用 在Release页面中下载打包好的程序和midi压缩包并解压。 双击运行“疯物之诗琴.exe”。 在原神中打开风物之诗琴,软件内输入

435 Jan 04, 2023
Rhasspy 673 Dec 28, 2022
Natural Language Processing Tasks and Examples.

Natural Language Processing Tasks and Examples With the advancement of A.I. technology in recent years, natural language processing technology has bee

Soohwan Kim 53 Dec 20, 2022
Continuously update some NLP practice based on different tasks.

NLP_practice We will continuously update some NLP practice based on different tasks. prerequisites Software pytorch = 1.10 torchtext = 0.11.0 sklear

0 Jan 05, 2022
IMS-Toucan is a toolkit to train state-of-the-art Speech Synthesis models

IMS-Toucan is a toolkit to train state-of-the-art Speech Synthesis models. Everything is pure Python and PyTorch based to keep it as simple and beginner-friendly, yet powerful as possible.

Digital Phonetics at the University of Stuttgart 247 Jan 05, 2023
auto_code_complete is a auto word-completetion program which allows you to customize it on your need

auto_code_complete v1.3 purpose and usage auto_code_complete is a auto word-completetion program which allows you to customize it on your needs. the m

RUO 2 Feb 22, 2022
Yodatranslator is a simple translator English to Yoda-language

yodatranslator Overview yodatranslator is a simple translator English to Yoda-language. Project is created for educational purposes. It is intended to

1 Nov 11, 2021
Задания КЕГЭ по информатике 2021 на Python

КЕГЭ 2021 на Python В этом репозитории мои решения типовых заданий КЕГЭ по информатике в 2021 году, БЕСПЛАТНО! Задания Взяты с https://inf-ege.sdamgia

8 Oct 13, 2022
TextFlint is a multilingual robustness evaluation platform for natural language processing tasks,

TextFlint is a multilingual robustness evaluation platform for natural language processing tasks, which unifies general text transformation, task-specific transformation, adversarial attack, sub-popu

TextFlint 587 Dec 20, 2022
Contains analysis of trends from Fitbit Dataset (source: Kaggle) to see how the trends can be applied to Bellabeat customers and Bellabeat products

Contains analysis of trends from Fitbit Dataset (source: Kaggle) to see how the trends can be applied to Bellabeat customers and Bellabeat products.

Leah Pathan Khan 2 Jan 12, 2022
Multi Task Vision and Language

12-in-1: Multi-Task Vision and Language Representation Learning Please cite the following if you use this code. Code and pre-trained models for 12-in-

Meta Research 711 Jan 08, 2023
Word2Wave: a framework for generating short audio samples from a text prompt using WaveGAN and COALA.

Word2Wave is a simple method for text-controlled GAN audio generation. You can either follow the setup instructions below and use the source code and CLI provided in this repo or you can have a play

Ilaria Manco 91 Dec 23, 2022
Code for the paper: Sequence-to-Sequence Learning with Latent Neural Grammars

Code for the paper: Sequence-to-Sequence Learning with Latent Neural Grammars

Yoon Kim 43 Dec 23, 2022
Simple and efficient RevNet-Library with DeepSpeed support

RevLib Simple and efficient RevNet-Library with DeepSpeed support Features Half the constant memory usage and faster than RevNet libraries Less memory

Lucas Nestler 112 Dec 05, 2022
:hot_pepper: R²SQL: "Dynamic Hybrid Relation Network for Cross-Domain Context-Dependent Semantic Parsing." (AAAI 2021)

R²SQL The PyTorch implementation of paper Dynamic Hybrid Relation Network for Cross-Domain Context-Dependent Semantic Parsing. (AAAI 2021) Requirement

huybery 60 Dec 31, 2022
This is a modification of the OpenAI-CLIP repository of moein-shariatnia

This is a modification of the OpenAI-CLIP repository of moein-shariatnia

Sangwon Beak 2 Mar 04, 2022
A natural language processing model for sequential sentence classification in medical abstracts.

NLP PubMed Medical Research Paper Abstract (Randomized Controlled Trial) A natural language processing model for sequential sentence classification in

Hemanth Chandran 1 Jan 17, 2022
2021 2학기 데이터크롤링 기말프로젝트

공지 주제 웹 크롤링을 이용한 취업 공고 스케줄러 스케줄 주제 정하기 코딩하기 핵심 코드 설명 + 피피티 구조 구상 // 12/4 토 피피티 + 스크립트(대본) 제작 + 녹화 // ~ 12/10 ~ 12/11 금~토 영상 편집 // ~12/11 토 웹크롤러 사람인_평균

Choi Eun Jeong 2 Aug 16, 2022