Publication describing 3 ML examples at NSLS-II and interfacing into Bluesky

Overview

Machine learning enabling high-throughput and remote operations at large-scale user facilities.

Overview

This repository contains the source code and examples for recreating the publication at arXiv:2201.03550.

Abstract

Imaging, scattering, and spectroscopy are fundamental in understanding and discovering new functional materials. Contemporary innovations in automation and experimental techniques have led to these measurements being performed much faster and with higher resolution, thus producing vast amounts of data for analysis. These innovations are particularly pronounced at user facilities and synchrotron light sources. Machine learning (ML) methods are regularly developed to process and interpret large datasets in real-time with measurements. However, there remain conceptual barriers to entry for the facility general user community, whom often lack expertise in ML, and technical barriers for deploying ML models. Herein, we demonstrate a variety of archetypal ML models for on-the-fly analysis at multiple beamlines at the National Synchrotron Light Source II (NSLS-II). We describe these examples instructively, with a focus on integrating the models into existing experimental workflows, such that the reader can easily include their own ML techniques into experiments at NSLS-II or facilities with a common infrastructure. The framework presented here shows how with little effort, diverse ML models operate in conjunction with feedback loops via integration into the existing Bluesky Suite for experimental orchestration and data management.

Explanation of Examples

As with all things at a user facility, each model is trained or set-up according to the needs of the user and their science. What is consistent across all AI agents, is their final communication paradigm. The agent loads and stores the model and/or necessary data, and has at minimum the following methods.

  • tell : tell the agent about some new data
  • report : construct a report (message, visualization, etc.) about the data
  • ask : ask the agent what to do next (for more see bluesky-adaptive)

Unsupervised learning (Non-negative matrix factorization)

The NMF companion agent keeps a constant cache of data to perform the reduction on. We treat these data as dependent variables, with independent variables coming fom the experiment. In the case study presented, the independent variables are temperature measurements, and the dependent variables are the 1-d spectra. Each call to report updates the decomposition using the full dataset, and updates the plots in the visualization.

The NMF companion agent is wrapped in a filesystem watcher, DirectoryAgent, which monitors a directory periodically. If there is new data in the target directory, the DirectoryAgent tells the NMF companion about the new data, and triggers a new report.

The construction of these objects, training, and visualization are all contained in the run_unsupervised file and mirrored in the corresponding notebook.

Anomaly detection

The model attributes a new observation to either normal or anomalous time series by comparing it to a large courpus of data collected at the beamline over an extended period of time. The development and updating of the model is done offline. Due to the nature of exparimental measurements, anomalous observatons may constitute a sizable portion of data withing a single collection period. Thus, a labeling of the data is required prior to model training. Once the model is trained it is saved as a binary file and loaded each time when AnomalyAgent is initialized.

A set of features devired from the original raw data, allowing the model to process time series of arbitary length.

The training can be found at run_anomaly.py with example deployment infrastructure at deploy_anomaly.py.

Supervised learning (Failure Classification)

The classifications of failures involves training the models entirely offline. This allows for robust model selection and specific deployment. A suite of models from scikit-learn are trained and tested, with the most promising model chosen to deploy. Since the models are lightweight, we re-train them at each instantiation during deployment with the most current dataset. For deep learning models, it would be appropriate to save and version the weights of a model, can construct the model at instantiation and load the weights.

The training can be found at run_supervised.py with example deployment infrastructure at deploy_supervised.py. How this is implemented at the BMM beamline can be found concisely here, where a wrapper agent does pointwise evaluation on UIDs of a document stream, using the ClassificationAgent's tell--report interface.

System Requirements

Hardware Requirements

Software Requirements

OS Requirements

This package has been tested exclusively on Linux operating systems.

  • RHEL 8.3
  • Ubuntu 18.04
  • PopOS 20.04

Python dependencies

  • numpy
  • matplotlib
  • scikit-learn
  • ipython

Getting Started

Installation guide

Install from github:

$ python3 -m venv pub_env
$ source pub_env/bin/activate
Owner
BNL
Brookhaven National Laboratory
BNL
HyperaPy: An automatic hyperparameter optimization framework ⚡🚀

hyperpy HyperPy: An automatic hyperparameter optimization framework Description HyperPy: Library for automatic hyperparameter optimization. Build on t

Sergio Mora 7 Sep 06, 2022
Code release for NeuS

NeuS We present a novel neural surface reconstruction method, called NeuS, for reconstructing objects and scenes with high fidelity from 2D image inpu

Peng Wang 813 Jan 04, 2023
Text to image synthesis using thought vectors

Text To Image Synthesis Using Thought Vectors This is an experimental tensorflow implementation of synthesizing images from captions using Skip Though

Paarth Neekhara 2.1k Jan 05, 2023
WPPNets: Unsupervised CNN Training with Wasserstein Patch Priors for Image Superresolution

WPPNets: Unsupervised CNN Training with Wasserstein Patch Priors for Image Superresolution This code belongs to the paper [1] available at https://arx

Fabian Altekrueger 5 Jun 02, 2022
A unified 3D Transformer Pipeline for visual synthesis

Overview This is the official repo for the paper: "NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion". NÜWA is a unified multimodal

Microsoft 2.6k Jan 03, 2023
Chinese named entity recognization with BiLSTM using Keras

Chinese named entity recognization (Bilstm with Keras) Project Structure ./ ├── README.md ├── data │   ├── README.md │   ├── data 数据集 │   │   ├─

1 Dec 17, 2021
Code of 3D Shape Variational Autoencoder Latent Disentanglement via Mini-Batch Feature Swapping for Bodies and Faces

3D Shape Variational Autoencoder Latent Disentanglement via Mini-Batch Feature Swapping for Bodies and Faces Installation After cloning the repo open

37 Dec 03, 2022
Source code of "Hold me tight! Influence of discriminative features on deep network boundaries"

Hold me tight! Influence of discriminative features on deep network boundaries This is the source code to reproduce the experiments of the NeurIPS 202

EPFL LTS4 19 Dec 10, 2021
Prososdy Morph: A python library for manipulating pitch and duration in an algorithmic way, for resynthesizing speech.

ProMo (Prosody Morph) Questions? Comments? Feedback? Chat with us on gitter! A library for manipulating pitch and duration in an algorithmic way, for

Tim 71 Jan 02, 2023
Implementation of "Unsupervised Domain Adaptive 3D Detection with Multi-Level Consistency"

Unsupervised Domain Adaptive 3D Detection with Multi-Level Consistency (ICCV2021) Paper Link: https://arxiv.org/abs/2107.11355 This implementation bui

32 Nov 17, 2022
Differentiable Prompt Makes Pre-trained Language Models Better Few-shot Learners

DART Implementation for ICLR2022 paper Differentiable Prompt Makes Pre-trained Language Models Better Few-shot Learners. Environment

ZJUNLP 83 Dec 27, 2022
Official Repsoitory for "Activate or Not: Learning Customized Activation." [CVPR 2021]

CVPR 2021 | Activate or Not: Learning Customized Activation. This repository contains the official Pytorch implementation of the paper Activate or Not

184 Dec 27, 2022
Lane assist for ETS2, built with the ultra-fast-lane-detection model.

Euro-Truck-Simulator-2-Lane-Assist Lane assist for ETS2, built with the ultra-fast-lane-detection model. This project was made possible by the amazing

36 Jan 05, 2023
A Fast Sequence Transducer Implementation with PyTorch Bindings

transducer A Fast Sequence Transducer Implementation with PyTorch Bindings. The corresponding publication is Sequence Transduction with Recurrent Neur

Awni Hannun 184 Dec 18, 2022
这个开源项目主要是对经典的时间序列预测算法论文进行复现,模型主要参考自GluonTS,框架主要参考自Informer

Time Series Research with Torch 这个开源项目主要是对经典的时间序列预测算法论文进行复现,模型主要参考自GluonTS,框架主要参考自Informer。 建立原因 相较于mxnet和TF,Torch框架中的神经网络层需要提前指定输入维度: # 建立线性层 TensorF

Chi Zhang 85 Dec 29, 2022
This is the code of NeurIPS'21 paper "Towards Enabling Meta-Learning from Target Models".

ST This is the code of NeurIPS 2021 paper "Towards Enabling Meta-Learning from Target Models". If you use any content of this repo for your work, plea

Su Lu 7 Dec 06, 2022
Powerful and efficient Computer Vision Annotation Tool (CVAT)

Computer Vision Annotation Tool (CVAT) CVAT is free, online, interactive video and image annotation tool for computer vision. It is being used by our

OpenVINO Toolkit 8.6k Jan 01, 2023
Implementation of character based convolutional neural network

Character Based CNN This repo contains a PyTorch implementation of a character-level convolutional neural network for text classification. The model a

Ahmed BESBES 248 Nov 21, 2022
Open source Python module for computer vision

About PCV PCV is a pure Python library for computer vision based on the book "Programming Computer Vision with Python" by Jan Erik Solem. More details

Jan Erik Solem 1.9k Jan 06, 2023
Dense Passage Retriever - is a set of tools and models for open domain Q&A task.

Dense Passage Retrieval Dense Passage Retrieval (DPR) - is a set of tools and models for state-of-the-art open-domain Q&A research. It is based on the

Meta Research 1.1k Jan 03, 2023