We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: ParticleAugment: Sampling-Based Data Augmentation

Abstract: We present an automated data augmentation approach for image classification. We formulate the problem as Monte Carlo sampling where our goal is to approximate the optimal augmentation policies. We propose a particle filtering scheme for the policy search where the probability of applying a set of augmentation operations forms the state of the filter. We measure the policy performance based on the loss function difference between a reference and the actual model, which we afterwards use to re-weight the particles and finally update the policy. In our experiments, we show that our formulation for automated augmentation reaches promising results on CIFAR-10, CIFAR-100, and ImageNet datasets using the standard network architectures for this problem. By comparing with the related work, our method reaches a balance between the computational cost of policy search and the model performance. Our code will be made publicly available.
Comments: 8 pages
Subjects: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2106.08693 [cs.LG]
  (or arXiv:2106.08693v3 [cs.LG] for this version)

Submission history

From: Alexander Tsaregorodtsev [view email]
[v1] Wed, 16 Jun 2021 10:56:02 GMT (26kb)
[v2] Tue, 24 Aug 2021 10:58:54 GMT (42kb)
[v3] Fri, 15 Oct 2021 14:43:07 GMT (120kb)

Link back to: arXiv, form interface, contact.