Summary of Model Parallel Training and Transfer Learning For Convolutional Neural Networks by Domain Decomposition, By Axel Klawonn and Martin Lanser and Janine Weber
Model Parallel Training and Transfer Learning for Convolutional Neural Networks by Domain Decomposition
by Axel Klawonn, Martin Lanser, Janine Weber
First submitted to arxiv on: 26 Aug 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG); Numerical Analysis (math.NA)
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 proposed novel model parallel CNN architecture is loosely inspired by domain decomposition and consists of decomposing input data into smaller subimages. Local CNNs with proportionally smaller numbers of parameters are trained in parallel, and the resulting local classifications are aggregated by a dense feedforward neural network (DNN). The paper compares this architecture to less costly alternatives for combining local classifications and investigates its performance when trained as one coherent model or using transfer learning. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Deep learning models have become incredibly successful in image processing tasks. To train these complex models efficiently, parallelization strategies are necessary. A new CNN architecture was proposed that breaks down input data into smaller pieces and trains local models simultaneously. This paper compares this approach to others for combining the local results and explores how it performs when trained as one model or using pre-trained models. |
Keywords
» Artificial intelligence » Cnn » Deep learning » Neural network » Transfer learning
