Research map / Learning theory and algorithms / Matrix and tensor methods
Eliminating the Invariance on the Loss Landscape of Linear Autoencoders
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| replace MSE with nested sum-of-p squared error loss L (sum over i of ||Y - A I_i,p B X||^2) | linear autoencoder decoder · Synthetic Gaussian data (n=1000, p=100) | fraction of principal directions correctly identified (RatioTP): from 0 to 100% | Helped |
| use nested loss L instead of standard MSE | linear autoencoder · MNIST (p=100) | reconstruction quality using first 10 decoder columns | Helped |
Closest papers by meaning
- Loss Landscapes of Regularized Linear Autoencoders (ICML 2019 · similarity 0.90)
- Be a Goldfish: Forgetting Bad Conditioning in Sparse Linear Regression via Variational Autoencoders (ICML 2025 · similarity 0.88)
- A Statistical Analysis of Wasserstein Autoencoders for Intrinsically Low-dimensional Data (ICLR 2024 · similarity 0.87)
- Sparse Autoencoders, Again? (ICML 2025 · similarity 0.85)
- Compression of Structured Data with Autoencoders: Provable Benefit of Nonlinearities and Depth (ICML 2024 · similarity 0.85)
- SASA: Subspace-Aware Sparse Autoencoders for Effective Mechanistic Interpretability (NeurIPS 2026 · similarity 0.85)
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.