Scaling Backwards: Minimal Synthetic Pre-training?

• URL: http://arxiv.org/abs/2408.00677v2
• Date: Sat, 3 Aug 2024 04:29:58 GMT
• Title: Scaling Backwards: Minimal Synthetic Pre-training?
• Authors: Ryo Nakamura, Ryu Tadokoro, Ryosuke Yamada, Yuki M. Asano, Iro Laina, Christian Rupprecht, Nakamasa Inoue, Rio Yokota, Hirokatsu Kataoka,
• Abstract summary: We show that pre-training is effective even with minimal synthetic images. We find that a substantial reduction of synthetic images from 1k to 1 can even lead to an increase in pre-training performance. We extend our method from synthetic images to real images to see if a single real image can show similar pre-training effect.
• Score: 52.78699562832907
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Pre-training and transfer learning are an important building block of current computer vision systems. While pre-training is usually performed on large real-world image datasets, in this paper we ask whether this is truly necessary. To this end, we search for a minimal, purely synthetic pre-training dataset that allows us to achieve performance similar to the 1 million images of ImageNet-1k. We construct such a dataset from a single fractal with perturbations. With this, we contribute three main findings. (i) We show that pre-training is effective even with minimal synthetic images, with performance on par with large-scale pre-training datasets like ImageNet-1k for full fine-tuning. (ii) We investigate the single parameter with which we construct artificial categories for our dataset. We find that while the shape differences can be indistinguishable to humans, they are crucial for obtaining strong performances. (iii) Finally, we investigate the minimal requirements for successful pre-training. Surprisingly, we find that a substantial reduction of synthetic images from 1k to 1 can even lead to an increase in pre-training performance, a motivation to further investigate ''scaling backwards''. Finally, we extend our method from synthetic images to real images to see if a single real image can show similar pre-training effect through shape augmentation. We find that the use of grayscale images and affine transformations allows even real images to ''scale backwards''.

Related papers

• Scaling Laws of Synthetic Images for Model Training ... for Now [54.43596959598466]
  We study the scaling laws of synthetic images generated by state of the art text-to-image models. We observe that synthetic images demonstrate a scaling trend similar to, but slightly less effective than, real images in CLIP training.
  arXiv  Detail & Related papers  (2023-12-07T18:59:59Z)
• Synthetic Image Data for Deep Learning [0.294944680995069]
  Realistic synthetic image data rendered from 3D models can be used to augment image sets and train image classification semantic segmentation models. We show how high quality physically-based rendering and domain randomization can efficiently create a large synthetic dataset based on production 3D CAD models of a real vehicle.
  arXiv  Detail & Related papers  (2022-12-12T20:28:13Z)
• Replacing Labeled Real-image Datasets with Auto-generated Contours [20.234550996148748]
  We show that formula-driven supervised learning can match or even exceed that of ImageNet-21k without the use of real images. Images generated by formulas avoid the privacy/copyright issues, labeling cost and errors, and biases that real images suffer from.
  arXiv  Detail & Related papers  (2022-06-18T06:43:38Z)
• On Efficient Transformer and Image Pre-training for Low-level Vision [74.22436001426517]
  Pre-training has marked numerous state of the arts in high-level computer vision. We present an in-depth study of image pre-training. We find pre-training plays strikingly different roles in low-level tasks.
  arXiv  Detail & Related papers  (2021-12-19T15:50:48Z)
• Task2Sim : Towards Effective Pre-training and Transfer from Synthetic Data [74.66568380558172]
  We study the transferability of pre-trained models based on synthetic data generated by graphics simulators to downstream tasks. We introduce Task2Sim, a unified model mapping downstream task representations to optimal simulation parameters. It learns this mapping by training to find the set of best parameters on a set of "seen" tasks. Once trained, it can then be used to predict best simulation parameters for novel "unseen" tasks in one shot.
  arXiv  Detail & Related papers  (2021-11-30T19:25:27Z)
• Is Deep Image Prior in Need of a Good Education? [57.3399060347311]
  Deep image prior was introduced as an effective prior for image reconstruction. Despite its impressive reconstructive properties, the approach is slow when compared to learned or traditional reconstruction techniques. We develop a two-stage learning paradigm to address the computational challenge.
  arXiv  Detail & Related papers  (2021-11-23T15:08:26Z)
• Intrinsic Autoencoders for Joint Neural Rendering and Intrinsic Image Decomposition [67.9464567157846]
  We propose an autoencoder for joint generation of realistic images from synthetic 3D models while simultaneously decomposing real images into their intrinsic shape and appearance properties. Our experiments confirm that a joint treatment of rendering and decomposition is indeed beneficial and that our approach outperforms state-of-the-art image-to-image translation baselines both qualitatively and quantitatively.
  arXiv  Detail & Related papers  (2020-06-29T12:53:58Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.