Summary of Deep Generative Domain Adaptation with Temporal Attention For Cross-user Activity Recognition, by Xiaozhou Ye et al.
Deep Generative Domain Adaptation with Temporal Attention for Cross-User Activity Recognition
by Xiaozhou Ye, Kevin I-Kai Wang
First submitted to arxiv on: 12 Mar 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Human-Computer Interaction (cs.HC)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper presents a novel approach to domain adaptation in Human Activity Recognition (HAR) tasks, particularly in cross-user scenarios where data distribution discrepancies are common. The Deep Generative Domain Adaptation with Temporal Attention (DGDATA) method addresses this challenge by recognizing and integrating temporal relations within time series data during the domain adaptation process. By combining generative models with a Temporal Relation Attention mechanism, DGDATA improves classification performance in cross-user HAR tasks. The proposed method is evaluated on three public sensor-based HAR datasets, demonstrating its efficacy for various scenarios and applications. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about a new way to help machines recognize human activities when the data used to train them comes from different people or situations. Normally, these types of machine learning problems assume that all the data is similar and independent. But in real-life situations, this isn’t always true. For example, when recognizing activities using sensors, the same activity might be performed differently by different people. The paper presents a new approach called DGDATA (Deep Generative Domain Adaptation with Temporal Attention) to address these challenges. This method uses both generative models and attention mechanisms to better recognize activities across different people and situations. |
Keywords
» Artificial intelligence » Activity recognition » Attention » Classification » Domain adaptation » Machine learning » Time series
