vc1492a / tidd

An approach for detecting tsunamis using anomaly detection anomalies on sTec d/dt data from orbiting GPS satellites.

vc1492a/tidd Issues

List of TIDs for future work
Closed a year ago24
Create a diagram showing the relationships of all the objects
Closed a year ago7
Parallel Training running into problems
Closed 3 years ago8
Examine variability in dependent variable across beginning, middle, end of prediction periods and quantify that variability
Closed 3 years ago1
Move model preparation and training code to source, refactor notebooks
Closed 3 years ago10
Improve speed of dataset (image) generation code
Closed 3 years ago1
Update S3 files to include the labeled data
Closed 3 years ago2
Animation of image files for anomalous classes
Closed 3 years ago4
Update readme.md with details of image-based approach
Closed 3 years ago1
Color QQ-Plot observations according to the sod
Closed 3 years ago2
Use all days for validation not just day of earthquake
Closed 3 years ago1
Write test for rescale values. Make sure it can take pandas series object.
Closed 3 years ago1
Update the acknowledgements in the readme.md
Closed 3 years ago1
Deal with maui__G20, pgf1__G08, pgf5__G20 in the validation set
Closed 4 years ago2
Construct simulated real-world experiment
Closed 4 years ago23
Fine-tune early stopping
Closed 4 years ago
Fine-tune the learning rate scheduler
Closed 4 years ago
Remove history.csv and models directory from image tar datasets
Closed 4 years ago2
Revisit inclusion of missing indicator in experiments
Closed 4 years ago1
Examine and quantify the absolute error pre and post event (disturbance in data) for earthquake periods
Closed 4 years ago1
Examine variability in absolute errors across prediction periods and quantify that variability
Closed 4 years ago1
Submit IEEE Paper
Closed 4 years ago1
Review Draft IEEE Paper
Closed 4 years ago1
Overleaf document for paper
Closed 4 years ago2
Update data documentation and load image data to AWS S3
Closed 4 years ago10
Select an architecture for experiments
Closed 4 years ago6
Address class imbalance
Closed 4 years ago6
Programmatically calculate metrics after model training
Closed 4 years ago8
Update hyperdash logging to include additional metrics and parameters
Closed 4 years ago2
Set aside N tracks for "real-world" experiment
Closed 4 years ago5
Model.find_appropriate_lr function not returning a non-negative value
Closed 4 years ago4
Validate training and testing data
Closed 4 years ago4
Clean and migrate code to src folder
Closed 4 years ago3
Transition to FastAI Version 2
Closed 4 years ago5
Add hyperdash logging into pipeline
Closed 4 years ago3
Develop / employ appropriate model error (residual) handling approach
Closed 4 years ago6
setup group email
Closed 4 years ago2
Master to Main
Closed 4 years ago
Add instrumentation for modeling / experiment runs (time needed for training, architecture used, etc.)
Closed 4 years ago5
Move data hosting to AWS S3 bucket and add data directory to gitignore
Closed 4 years ago2
Identify (based on literature / analysis) typical length (or distribution) of the length of the disturbances in the signal
Closed 4 years ago3
Add model training and testing functionality with metrics
Closed 4 years ago3
Speed up read of data into memory for model training and analysis
Closed 4 years ago2
Explore adding additional pre-earthquake data to Hawaii experiment
Closed 4 years ago3
CUDA out of memory for Tesla P4
Closed 4 years ago4
gpustat requirements in repo
Closed 4 years ago2
Improve speed of tests by reducing test data sizes
Closed 4 years ago7
Ask Dr. Bortnik about using UCLA resources for computing and get environment specifications
Closed 4 years ago10
Determine performance metrics to include in experiments
Closed 4 years ago3
Explore and resolve missing data
Closed 4 years ago4