AutoJoin: Efficient Adversarial Training against Gradient-Free Perturbations for Robust Maneuvering via Denoising Autoencoder and Joint Learning
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2022
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866929574858719232 |
|---|---|
| author | Villarreal, Michael Poudel, Bibek Wickman, Ryan Shen, Yu Li, Weizi |
| author_facet | Villarreal, Michael Poudel, Bibek Wickman, Ryan Shen, Yu Li, Weizi |
| contents | With the growing use of machine learning algorithms and ubiquitous sensors, many `perception-to-control' systems are being developed and deployed. To ensure their trustworthiness, improving their robustness through adversarial training is one potential approach. We propose a gradient-free adversarial training technique, named AutoJoin, to effectively and efficiently produce robust models for image-based maneuvering. Compared to other state-of-the-art methods with testing on over 5M images, AutoJoin achieves significant performance increases up to the 40% range against perturbations while improving on clean performance up to 300%. AutoJoin is also highly efficient, saving up to 86% time per training epoch and 90% training data over other state-of-the-art techniques. The core idea of AutoJoin is to use a decoder attachment to the original regression model creating a denoising autoencoder within the architecture. This architecture allows the tasks `maneuvering' and `denoising sensor input' to be jointly learnt and reinforce each other's performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2205_10933 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | AutoJoin: Efficient Adversarial Training against Gradient-Free Perturbations for Robust Maneuvering via Denoising Autoencoder and Joint Learning Villarreal, Michael Poudel, Bibek Wickman, Ryan Shen, Yu Li, Weizi Machine Learning Computer Vision and Pattern Recognition Robotics With the growing use of machine learning algorithms and ubiquitous sensors, many `perception-to-control' systems are being developed and deployed. To ensure their trustworthiness, improving their robustness through adversarial training is one potential approach. We propose a gradient-free adversarial training technique, named AutoJoin, to effectively and efficiently produce robust models for image-based maneuvering. Compared to other state-of-the-art methods with testing on over 5M images, AutoJoin achieves significant performance increases up to the 40% range against perturbations while improving on clean performance up to 300%. AutoJoin is also highly efficient, saving up to 86% time per training epoch and 90% training data over other state-of-the-art techniques. The core idea of AutoJoin is to use a decoder attachment to the original regression model creating a denoising autoencoder within the architecture. This architecture allows the tasks `maneuvering' and `denoising sensor input' to be jointly learnt and reinforce each other's performance. |
| title | AutoJoin: Efficient Adversarial Training against Gradient-Free Perturbations for Robust Maneuvering via Denoising Autoencoder and Joint Learning |
| topic | Machine Learning Computer Vision and Pattern Recognition Robotics |
| url | https://arxiv.org/abs/2205.10933 |