Repo for paper "Dynamic Placement of Rapidly Deployable Mobile Sensor Robots Using Machine Learning and Expected Value of Information"

Last update: Jul 01, 2021

Related tags

Deep Learning EVSIvsLSTM

Overview

Repo for paper "Dynamic Placement of Rapidly Deployable Mobile Sensor Robots Using Machine Learning and Expected Value of Information"

Notes

I probably still left some absolute path in the project. When running a script, please check the paths used for loading dataset, saving models and etc.

Repo Structure

The section contains the structure graph of this project and some simple descriptions of folders

For more detailed description of each script, please refer to the README inside each folder.
Every item with . extension is a file/script. Items without . extension are folders.
Folders like dataset, models, backup are not actually empty. But because they usually hold fairly large files/datasets I decided to not upload the content directly to github (.gitignore are left in those folders as placeholders). Please contact me directly if you need those files.

│   env.yml: environment file (under Windows 10) for Conda. Use this to generate a working environment
│
├───evsi: scripts/data related to EVSI portion of this project
│   │   get_EVSI.ipynb
│   │   get_models.ipynb
│   │   ranking_and_correlation.ipynb
│   │   sensitivity_analysis.ipynb
│   │   training_and_evsi_fs.ipynb
│   │
│   ├───backup: results of each run of `training_and_evsi_fs.ipynb`.
│   ├───dataset: raw data of the TE dataset
│   ├───log: relevant metrics generated after the current run
│   │       acc.csv
│   │       acc_improvement.csv
│   │       sensitivity_analysis.csv
│   │       sensor_selection.csv
│   │
│   └───models: frozen LSTM models saved after the current run
│       ├───evsi: models trained for EVSI purpose
│       └───ml: models trained for forward stepwise selection purpose
|
└───ml
    │   LSTM_RandomForest.ipynb
    │   LSTM_workflow.ipynb
    │   README.md
    │   visulization.ipynb
    │
    ├───dataset: raw data of the TE dataset
    ├───models: frozen LSTM models that are used to pick the top 10 impactful features
    └───plots: plots generated to demonstrate the 10 most impactful features
            test_advantage.png
            validation_advantage.png

Dataset

The dataset used in this project is the Tennessee Eastman Process Simulation Data presented here

More Info

For more information about this project, for example, the structure of the dataset, please refer to this document

Repo for paper "Dynamic Placement of Rapidly Deployable Mobile Sensor Robots Using Machine Learning and Expected Value of Information"

Related tags

Overview

Notes

Repo Structure

Dataset

More Info

Owner

Berkeley Expert System Technologies Lab

Code repository for Self-supervised Structure-sensitive Learning, CVPR'17

BirdCLEF 2021 - Birdcall Identification 4th place solution

PhysCap: Physically Plausible Monocular 3D Motion Capture in Real Time

QICK: Quantum Instrumentation Control Kit

Text-to-Music Retrieval using Pre-defined/Data-driven Emotion Embeddings

City-Scale Multi-Camera Vehicle Tracking Guided by Crossroad Zones Code

DCA - Official Python implementation of Delaunay Component Analysis algorithm

A hybrid framework (neural mass model + ML) for SC-to-FC prediction

Human Activity Recognition example using TensorFlow on smartphone sensors dataset and an LSTM RNN. Classifying the type of movement amongst six activity categories - Guillaume Chevalier

Real-Time-Student-Attendence-System - Real Time Student Attendence System

Public implementation of "Learning from Suboptimal Demonstration via Self-Supervised Reward Regression" from CoRL'21

Speech Enhancement Generative Adversarial Network Based on Asymmetric AutoEncoder

VolumeGAN - 3D-aware Image Synthesis via Learning Structural and Textural Representations

Official implementation for ICDAR 2021 paper "Handwritten Mathematical Expression Recognition with Bidirectionally Trained Transformer"

scikit-learn inspired API for CRFsuite

Moving Object Segmentation in 3D LiDAR Data: A Learning-based Approach Exploiting Sequential Data

My implementation of Image Inpainting - A deep learning Inpainting model

Cockpit is a visual and statistical debugger specifically designed for deep learning.

Implementation of "RaScaNet: Learning Tiny Models by Raster-Scanning Image" from CVPR 2021.

Learn the Deep Learning for Computer Vision in three steps: theory from base to SotA, code in PyTorch, and space-repetition with Anki