This repository contains a PyTorch implementation of the paper Learning to Assimilate in Chaotic Dynamical Systems.

Overview

Amortized Assimilation

This repository contains a PyTorch implementation of the paper Learning to Assimilate in Chaotic Dynamical Systems.

Abstract: The accuracy of simulation-based forecasting in chaotic systems is heavily dependent on high-quality estimates of the system state at the time the forecast is initialized. Data assimilation methods are used to infer these initial conditions by systematically combining noisy, incomplete observations and numerical models of system dynamics to produce effective estimation schemes. We introduce amortized assimilation, a framework for learning to assimilate in dynamical systems from sequences of noisy observations with no need for ground truth data. We motivate the framework by extending powerful results from self-supervised denoising to the dynamical systems setting through the use of differentiable simulation.

Installation

Requirements

This code can be memory heavy as each experiment unrolls at least 40 assimilation steps (which from a memory perspective is equivalent to a 40x deeper network plus whatever is needed for the simulation). Current settings are optimized to max out memory usage on a GTX1070 GPU. The easiest ways to tune memory usage are network width and ensemble size. Checkpointing could significantly improve memory utilization but is not currently implemented.

To install the dependencies, use the provided requirements.txt file:

pip install -r requirements.txt 

There is also a dependency on torchdiffeq. Instructions for installing torchdiffeq can be found at https://github.com/rtqichen/torchdiffeq, but are also copied below:

pip install git+https://github.com/rtqichen/torchdiffeq

To run the DA comparison models, you will need to install DAPPER. Instructions can be found here: https://github.com/nansencenter/DAPPER.

Installing this package

A setup.py file has been included for installation. Navigate to the home folder and run:

pip install -e . 

Run experiments

All experiments can be run from experiments/run_*.py. Default settings are those used in the paper. First navigate to the experiments directory then execute:

L96 Full Observations

python run_L96Conv.py --obs_conf full_obs

L96 Partial Observations (every fourth).

python run_L96Conv.py --obs_conf every_4th_dim_partial_obs

VL20 Partial

python run_VLConv.py --obs_conf every_4th_dim_partial_obs

KS Full

python run_KS.py 

Other modifications of interest might be to adjust the step size for the integrator (--step_size, default .1), observation error(--noise, default 1.), ensemble size (--m, default 10), or network width (--hidden_size, default 64 for conv). The L96 code also includes options for self-supervised and supervised analysis losses (ss_analysis, clean_analysis) used for creating Figure 6 from the paper. Custom observation operators can be created in the same style as those found in obs_configs.py.

Parameters for traditional DA approaches were tuned via grid search over smaller sequences. Those hyperparameters were then used for longer assimilation sequences.

To test a new architecture, you'll want to ensure it's obeying the same API as the models in models.py, but otherwise it should slot in without major issues.

Datasets

Code is included for generating the Lorenz 96, VL 20 and KS datasets. This can be found under amortized_assimilation/data_utils.py

References

DAPPER: Raanes, P. N., & others. (2018). nansencenter/DAPPER: Version 0.8. https://doi.org/10.5281/zenodo.2029296

Acknowledgements

This material is based upon work supported by the National Science Foundation under Grant No. 1835825. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.


If you found the code or ideas in this repository useful, please consider citing:

@article{mccabe2021l2assim,
  title={Learning to Assimilate in Chaotic Dynamical Systems},
  author={McCabe, Michael and Brown, Jed},
  journal={Advances in Neural Information Processing Systems},
  year={2021}
}
Nicholas Lee 3 Jan 09, 2022
Spectrum is an AI that uses machine learning to generate Rap song lyrics

Spectrum Spectrum is an AI that uses deep learning to generate rap song lyrics. View Demo Report Bug Request Feature Open In Colab About The Project S

39 Dec 16, 2022
An interpreter for RASP as described in the ICML 2021 paper "Thinking Like Transformers"

RASP Setup Mac or Linux Run ./setup.sh . It will create a python3 virtual environment and install the dependencies for RASP. It will also try to insta

141 Jan 03, 2023
Training RNNs as Fast as CNNs

News SRU++, a new SRU variant, is released. [tech report] [blog] The experimental code and SRU++ implementation are available on the dev branch which

ASAPP Research 2.1k Jan 01, 2023
Simplified interface for TensorFlow (mimicking Scikit Learn) for Deep Learning

SkFlow has been moved to Tensorflow. SkFlow has been moved to http://github.com/tensorflow/tensorflow into contrib folder specifically located here. T

3.2k Dec 29, 2022
PyTorch code for Composing Partial Differential Equations with Physics-Aware Neural Networks

FInite volume Neural Network (FINN) This repository contains the PyTorch code for models, training, and testing, and Python code for data generation t

Cognitive Modeling 20 Dec 18, 2022
Code and data for paper "Deep Photo Style Transfer"

deep-photo-styletransfer Code and data for paper "Deep Photo Style Transfer" Disclaimer This software is published for academic and non-commercial use

Fujun Luan 9.9k Dec 29, 2022
PyTorch deep learning projects made easy.

PyTorch Template Project PyTorch deep learning project made easy. PyTorch Template Project Requirements Features Folder Structure Usage Config file fo

Victor Huang 3.8k Jan 01, 2023
Hl classification bc - A Network-Based High-Level Data Classification Algorithm Using Betweenness Centrality

A Network-Based High-Level Data Classification Algorithm Using Betweenness Centr

Esteban Vilca 3 Dec 01, 2022
An open-source project for applying deep learning to medical scenarios

Auto Vaidya An open source solution for creating end-end web app for employing the power of deep learning in various clinical scenarios like implant d

Smaranjit Ghose 18 May 29, 2022
Keras udrl - Keras implementation of Upside Down Reinforcement Learning

keras_udrl Keras implementation of Upside Down Reinforcement Learning This is me

Eder Santana 7 Jan 24, 2022
Code for our paper "Interactive Analysis of CNN Robustness"

Perturber Code for our paper "Interactive Analysis of CNN Robustness" Datasets Feature visualizations: Google Drive Fine-tuning checkpoints as saved m

Stefan Sietzen 0 Aug 17, 2021
A python interface for training Reinforcement Learning bots to battle on pokemon showdown

The pokemon showdown Python environment A Python interface to create battling pokemon agents. poke-env offers an easy-to-use interface for creating ru

Haris Sahovic 184 Dec 30, 2022
Digan - Official PyTorch implementation of Generating Videos with Dynamics-aware Implicit Generative Adversarial Networks

DIGAN (ICLR 2022) Official PyTorch implementation of "Generating Videos with Dyn

Sihyun Yu 147 Dec 31, 2022
Official implementation of the NeurIPS'21 paper 'Conditional Generation Using Polynomial Expansions'.

Conditional Generation Using Polynomial Expansions Official implementation of the conditional image generation experiments as described on the NeurIPS

Grigoris 4 Aug 07, 2022
RP-GAN: Stable GAN Training with Random Projections

RP-GAN: Stable GAN Training with Random Projections This repository contains a reference implementation of the algorithm described in the paper: Behna

Ayan Chakrabarti 20 Sep 18, 2021
Codebase for Time-series Generative Adversarial Networks (TimeGAN)

Codebase for Time-series Generative Adversarial Networks (TimeGAN)

Jinsung Yoon 532 Dec 31, 2022
CVPRW 2021: How to calibrate your event camera

E2Calib: How to Calibrate Your Event Camera This repository contains code that implements video reconstruction from event data for calibration as desc

Robotics and Perception Group 104 Nov 16, 2022
[WACV 2020] Reducing Footskate in Human Motion Reconstruction with Ground Contact Constraints

Reducing Footskate in Human Motion Reconstruction with Ground Contact Constraints Official implementation for Reducing Footskate in Human Motion Recon

Virginia Tech Vision and Learning Lab 38 Nov 01, 2022
Meta Learning for Semi-Supervised Few-Shot Classification

few-shot-ssl-public Code for paper Meta-Learning for Semi-Supervised Few-Shot Classification. [arxiv] Dependencies cv2 numpy pandas python 2.7 / 3.5+

Mengye Ren 501 Jan 08, 2023