This is the code of AAAI'21 paper "Tailoring Embedding Function to Heterogeneous Few-Shot Tasks by Global and Local Feature Adaptors".
The code trains an LSTM-based residual model on human motion data and applies transfer learning to detect robotic joint faults. It preprocesses data, maps robot features to human-like patterns, and fine-tunes a model while freezing early layers. The optimized model is evaluated with class weighting, callbacks, and feature importance analysis.
The code trains an LSTM-based residual model on human motion data and applies transfer learning to detect robotic joint faults. It preprocesses data, maps robot features to human-like patterns, and fine-tunes a model while freezing early layers. The optimized model is evaluated with class weighting, callbacks, and feature importance analysis.
Add a description, image, and links to the feature-adaptation topic page so that developers can more easily learn about it.
To associate your repository with the feature-adaptation topic, visit your repo's landing page and select "manage topics."