Summary of Learning Multiresolution Matrix Factorization and Its Wavelet Networks on Graphs, by Truong Son Hy and Risi Kondor
Learning Multiresolution Matrix Factorization and its Wavelet Networks on Graphs
by Truong Son Hy, Risi Kondor
First submitted to arxiv on: 2 Nov 2021
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 This paper proposes a learnable version of Multiresolution Matrix Factorization (MMF), which is well-suited for modeling complex graphs with hierarchical structures. Unlike other fast matrix factorization algorithms, MMF does not assume low rank, making it particularly effective in this context. The challenge lies in finding the factorization itself, as existing greedy methods tend to be brittle. To overcome this, the authors combine reinforcement learning and Stiefel manifold optimization through backpropagating errors. The resulting wavelet basis outperforms prior MMF algorithms and enables robust deployment on standard learning tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper creates a new way of breaking down complex graphs into smaller parts using Multiresolution Matrix Factorization (MMF). Unlike other methods, MMF doesn’t assume the graph is simple, making it perfect for finding patterns in complicated data. The tricky part is figuring out how to do this efficiently, which existing methods struggle with. To solve this problem, the researchers use a combination of two powerful techniques: reinforcement learning and Stiefel manifold optimization. This new approach creates a better wavelet basis that can be used in many different applications. |
Keywords
* Artificial intelligence * Optimization * Reinforcement learning
