From bf78dd3ce050d94f87ea73e15337fb2e2e52447a Mon Sep 17 00:00:00 2001 From: Hassan Ismail Fawaz Date: Sat, 20 Jul 2024 11:39:18 +0000 Subject: [PATCH] add uda aep --- aep/06_domain_adaptation.md | 48 +++++++++++++++++++++++++++++++++++++ 1 file changed, 48 insertions(+) create mode 100644 aep/06_domain_adaptation.md diff --git a/aep/06_domain_adaptation.md b/aep/06_domain_adaptation.md new file mode 100644 index 0000000..3919c31 --- /dev/null +++ b/aep/06_domain_adaptation.md @@ -0,0 +1,48 @@ +# AEP UDA + +## Overview + +Introducing a benchmark module for evaluating Unsupervised Domain Adaptation (UDA) techniques for time series classification using various backbones with different techniques for doing domain adaptation. + +## Problem Statement and Use Cases + +UDA for time series data is underexplored despite its importance in fields like medicine, manufacturing, and earth observation. This AEP addresses this gap by providing access to UDA methods, helping improve models for unlabeled target data. + +### Use Cases +- Medical diagnostics +- Predictive maintenance in manufacturing +- Environmental monitoring +- Human activity recognition + +## Implementation + +1. Compile a dataset with source and target seperation. +2. Design a standardized evaluation framework. +3. Integrate neural network backbones like Inception. +4. Evaluate UDA techniques using the framework. + +## Example Code/Structure + +```python +# Load and evaluate a UDA technique on the benchmark dataset + +import benchmark_framework + +dataset = benchmark_framework.load_dataset(x_source,y_source,x_target,y_target) +uda_model = benchmark_framework.UDA_Model(backbone='Inception') +results = uda_model.evaluate(dataset) +print(results) +``` + +## Considerations and Alternatives + +- Ensure compatibility with existing UDA tools. +- Explore alternative neural network backbones and evaluation metrics. + +## Discussion + +N/A + +## References + +- [Deep Unsupervised Domain Adaptation for Time Series Classification: a Benchmark](https://arxiv.org/abs/2312.09857) \ No newline at end of file