Meta Learning Backpropagation And Improving It (VSML)

Overview

Meta Learning Backpropagation And Improving It (VSML)

This is research code for the NeurIPS 2021 publication Kirsch & Schmidhuber 2021.

Many concepts have been proposed for meta learning with neural networks (NNs), e.g., NNs that learn to reprogram fast weights, Hebbian plasticity, learned learning rules, and meta recurrent NNs. Our Variable Shared Meta Learning (VSML) unifies the above and demonstrates that simple weight-sharing and sparsity in an NN is sufficient to express powerful learning algorithms (LAs) in a reusable fashion. A simple implementation of VSML where the weights of a neural network are replaced by tiny LSTMs allows for implementing the backpropagation LA solely by running in forward-mode. It can even meta learn new LAs that differ from online backpropagation and generalize to datasets outside of the meta training distribution without explicit gradient calculation. Introspection reveals that our meta learned LAs learn through fast association in a way that is qualitatively different from gradient descent.

Installation

Create a virtual env

python3 -m venv venv
. venv/bin/activate

Install pip dependencies

pip3 install --upgrade pip wheel setuptools
pip3 install -r requirements.txt

Initialize weights and biases

wandb init

Inspect your results at https://wandb.ai/.

Run instructions

Non distributed

For any algorithm that does not require multiple workers.

python3 launch.py --config_files CONFIG_FILES --config arg1=val1 arg2=val2

Distributed

For any algorithm that does require multiple workers

GPU_COUNT=4 mpirun -n NUM_WORKERS python3 assign_gpu.py python3 launch.py

where NUM_WORKERS is the number of workers to run. The assign_gpu python script distributes the mpi workers evenly over the specified GPUs

Alternatively, specify the CUDA_VISIBLE_DEVICES instead of GPU_COUNT env variable:

CUDA_VISIBLE_DEVICES=0,2,3 mpirun -n NUM_WORKERS python3 assign_gpu.py python3 launch.py

Slurm-based cluster

Modify slurm/schedule.sh and slurm/job.sh to suit your environment.

bash slurm/schedule.sh --nodes=7 --ntasks-per-node=12 -- python3 launch.py --config_files CONFIG_FILES

If only a single worker is required (non-distributed), set --nodes=1 and --ntasks-per-node=1.

Remote (via ssh)

Modify ssh/schedule.sh to suit your environment. Requires gpustat in .local/bin/gpustat, via pip3 install --user gpustat. Also install tmux and mpirun.

bash ssh/schedule.sh --host HOST_NAME --nodes=7 --ntasks-per-node=12 -- python3 launch.py --config_files CONFIG_FILES

Example training runs

Section 4.2 Figure 6

VSML

slurm/schedule.py --nodes=128 --time 04:00:00 -- python3 launch.py --config_files configs/rand_proj.yaml

You can also try fewer nodes and use --config training.population_size=128. Or use backpropagation-based meta optimization --config_files configs/{rand_proj,backprop}.yaml.

Section 4.4 Figure 8

VSML

slurm/schedule.py --array=1-11 --nodes=128 --time 04:00:00 -- python3 launch.py --array configs/array/datasets.yaml

Meta RNN (Hochreiter 2001)

slurm/schedule.py --array=1-11 --nodes=32 --time 04:00:00 -- python3 launch.py --array configs/array/datasets.yaml --config_files configs/{metarnn,pad}.yaml --tags metarnn

Fast weight memory

slurm/schedule.py --array=1-11 --nodes=32 --time 04:00:00 -- python3 launch.py --array configs/array/datasets.yaml --config_files configs/{fwmemory,pad}.yaml --tags fwmemory

SGD

slurm/schedule.py --array=1-4 --nodes=2 --time 00:15:00 -- python3 launch.py --array configs/array/sgd.yaml --config_files configs/sgd.yaml --tags sgd

Hebbian

slurm/schedule.py --array=1-11 --nodes=32 --time 04:00:00 -- python3 launch.py --array configs/array/datasets.yaml --config_files configs/{hebbian,pad}.yaml --tags hebbian
Owner
Louis Kirsch
Building RL agents that meta-learn their own learning algorithm. Currently pursuing a PhD in AI at IDSIA with Jürgen Schmidhuber. Previous DeepMind intern.
Louis Kirsch
An LSTM for time-series classification

Update 10-April-2017 And now it works with Python3 and Tensorflow 1.1.0 Update 02-Jan-2017 I updated this repo. Now it works with Tensorflow 0.12. In

Rob Romijnders 391 Dec 27, 2022
PyTorch implementation of: Michieli U. and Zanuttigh P., "Continual Semantic Segmentation via Repulsion-Attraction of Sparse and Disentangled Latent Representations", CVPR 2021.

Continual Semantic Segmentation via Repulsion-Attraction of Sparse and Disentangled Latent Representations This is the official PyTorch implementation

Multimedia Technology and Telecommunication Lab 42 Nov 09, 2022
ViDT: An Efficient and Effective Fully Transformer-based Object Detector

ViDT: An Efficient and Effective Fully Transformer-based Object Detector by Hwanjun Song1, Deqing Sun2, Sanghyuk Chun1, Varun Jampani2, Dongyoon Han1,

NAVER AI 262 Dec 27, 2022
This folder contains the implementation of the multi-relational attribute propagation algorithm.

MrAP This folder contains the implementation of the multi-relational attribute propagation algorithm. It requires the package pytorch-scatter. Please

6 Dec 06, 2022
🐸STT integration examples

🐸 STT 0.9.x Examples These are various examples on how to use or integrate 🐸 STT using our packages. It is a good way to just try out 🐸 STT before

coqui 92 Dec 19, 2022
Drone-based Joint Density Map Estimation, Localization and Tracking with Space-Time Multi-Scale Attention Network

DroneCrowd Paper Detection, Tracking, and Counting Meets Drones in Crowds: A Benchmark. Introduction This paper proposes a space-time multi-scale atte

VisDrone 98 Nov 16, 2022
Implementation of MeMOT - Multi-Object Tracking with Memory - in Pytorch

MeMOT - Pytorch (wip) Implementation of MeMOT - Multi-Object Tracking with Memory - in Pytorch. This paper is just one in a line of work, but importan

Phil Wang 15 May 09, 2022
This repository contains the code for "SBEVNet: End-to-End Deep Stereo Layout Estimation" paper by Divam Gupta, Wei Pu, Trenton Tabor, Jeff Schneider

SBEVNet: End-to-End Deep Stereo Layout Estimation This repository contains the code for "SBEVNet: End-to-End Deep Stereo Layout Estimation" paper by D

Divam Gupta 19 Dec 17, 2022
This repo in the implementation of EMNLP'21 paper "SPARQLing Database Queries from Intermediate Question Decompositions" by Irina Saparina, Anton Osokin

SPARQLing Database Queries from Intermediate Question Decompositions This repo is the implementation of the following paper: SPARQLing Database Querie

Yandex Research 20 Dec 19, 2022
rliable is an open-source Python library for reliable evaluation, even with a handful of runs, on reinforcement learning and machine learnings benchmarks.

Open-source library for reliable evaluation on reinforcement learning and machine learning benchmarks. See NeurIPS 2021 oral for details.

Google Research 529 Jan 01, 2023
Repository for the Bias Benchmark for QA dataset.

BBQ Repository for the Bias Benchmark for QA dataset. Authors: Alicia Parrish, Angelica Chen, Nikita Nangia, Vishakh Padmakumar, Jason Phang, Jana Tho

ML² AT CILVR 18 Nov 18, 2022
Source code for PairNorm (ICLR 2020)

PairNorm Official pytorch source code for PairNorm paper (ICLR 2020) This code requires pytorch_geometric=1.3.2 usage For SGC, we use original PairNo

62 Dec 08, 2022
🐥A PyTorch implementation of OpenAI's finetuned transformer language model with a script to import the weights pre-trained by OpenAI

PyTorch implementation of OpenAI's Finetuned Transformer Language Model This is a PyTorch implementation of the TensorFlow code provided with OpenAI's

Hugging Face 1.4k Jan 05, 2023
Tgbox-bench - Simple TGBOX upload speed benchmark

TGBOX Benchmark This script will benchmark upload speed to TGBOX storage. Build

Non 1 Jan 09, 2022
Official codebase for "B-Pref: Benchmarking Preference-BasedReinforcement Learning" contains scripts to reproduce experiments.

B-Pref Official codebase for B-Pref: Benchmarking Preference-BasedReinforcement Learning contains scripts to reproduce experiments. Install conda env

48 Dec 20, 2022
ChebLieNet, a spectral graph neural network turned equivariant by Riemannian geometry on Lie groups.

ChebLieNet: Invariant spectral graph NNs turned equivariant by Riemannian geometry on Lie groups Hugo Aguettaz, Erik J. Bekkers, Michaël Defferrard We

haguettaz 12 Dec 10, 2022
A simple, high level, easy-to-use open source Computer Vision library for Python.

ZoomVision : Slicing Aid Detection A simple, high level, easy-to-use open source Computer Vision library for Python. Installation Installing dependenc

Nurettin Sinanoğlu 2 Mar 04, 2022
Spherical Confidence Learning for Face Recognition, accepted to CVPR2021.

Sphere Confidence Face (SCF) This repository contains the PyTorch implementation of Sphere Confidence Face (SCF) proposed in the CVPR2021 paper: Shen

Maths 70 Dec 09, 2022
[ICCV2021] IICNet: A Generic Framework for Reversible Image Conversion

IICNet - Invertible Image Conversion Net Official PyTorch Implementation for IICNet: A Generic Framework for Reversible Image Conversion (ICCV2021). D

felixcheng97 55 Dec 06, 2022
Contenido del curso Bases de datos del DCC PUC versión 2021-2

IIC2413 - Bases de Datos Tabla de contenidos Equipo Profesores Ayudantes Contenidos Calendario Evaluaciones Resumen de notas Foro Política de integrid

54 Nov 23, 2022