Research map / Optimization / Distributed and compressed
A Decentralized Framework for Kernel PCA With Projection Consensus Constraints
Fan He, Ruikai Yang, Lei Shi, Xiaolin Huang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add projection consensus constraints (Alg. 1) to local kPCA when local data are scarce | decentralized kPCA, 20-node network · MNIST (N_j from 40 to 300) | similarity to central kPCA solution: improvement is less significant when there are more data in local nodes | Improved |
| use consensus ADMM iterations instead of gathering all neighbors' raw data | decentralized kPCA, 20 nodes, 100 images each · MNIST (neighbors 2 to 12) | average similarity to central kPCA solution: with about 4 iterations, the similarity given by Alg. 1 exceeds that of obtaining all neighbor data | Improved |
| increase number of neighbors from 4 to 12 for Alg. 1 at convergence versus gathering 12 neighbors' data (ablation) | decentralized kPCA · MNIST (neighbors 2,4,6,8,10,12) | average similarity: similar to (|Omega_j|=2) or much better than (|Omega_j|=4,6,8,10,12) gathering 12 neighboring nodes | Improved |
| increase network size to 80 nodes (ablation) | decentralized kPCA · MNIST (100 images/node, 4 neighbors) | similarity and running time: similarity is kept above 0.912 | No clear effect |
| replace central kPCA by decentralized Alg. 1 | kernel PCA · MNIST running time | running time | Improved |
Closest papers by meaning
- Improved Analysis of the Accelerated Noisy Power Method with Applications to Decentralized PCA (ICML 2026 · similarity 0.86)
- DeEPCA: Decentralized Exact PCA with Linear Convergence Rate (JMLR 2021 · similarity 0.86)
- Communication-efficient Algorithms for Distributed Stochastic Principal Component Analysis (ICML 2017 · similarity 0.86)
- COKE: Communication-Censored Decentralized Kernel Learning (JMLR 2021 · similarity 0.85)
- Decentralized Learning for Overparameterized Problems: A Multi-Agent Kernel Approximation Approach (ICLR 2022 · similarity 0.85)
- Communication-Efficient Distributed PCA by Riemannian Optimization (ICML 2020 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
