AutoJoin: Efficient Adversarial Training against Gradient-Free Perturbations for Ro- bust Maneuvering via Denoising Autoencoder and Joint Learning
Primary Area: applications to robotics, autonomy, planning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Keywords: autonomous driving, gradient-free perturbations
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Abstract: 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 ad-
versarial training is one potential approach. We propose a gradient-free
adversarial training technique, named AutoJoin, to effectively and effi-
ciently 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%. Auto-
Join 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 archi-
tecture allows the tasks ‘maneuvering’ and ‘denoising sensor input’ to be
jointly learnt and reinforce each other’s performance. The project code is at
https://anonymous.4open.science/r/AutoJoin-FA13.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Supplementary Material: pdf
Submission Number: 4472
Loading