Research map / Probabilistic and scientific modeling / Molecules and proteins
Unsupervised Learning of Equivariant Structure from Sequences
Takeru Miyato, Masanori Koyama, Kenji Fukumizu
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 |
|---|---|---|---|
| Train an encoder-decoder to predict held-out future frames using a per-sequence linear latent transition | Sequential MNIST and Sequential MNIST-bg · Sequential MNIST and Sequential MNIST-bg | prediction quality | Improved |
| Use per-sequence neural transitions instead of a fitted linear latent transition (ablation) | MSP · Sequential MNIST, extrapolation | prediction quality | Hurt |
| Predict M* with a neural network rather than optimize it internally for each sequence (ablation) | MSP · Sequential MNIST, equivariance | equivariance error | Hurt |
| Constrain the latent transition to fixed 2x2 blocks (ablation) | MSP · Sequential MNIST, extrapolation | prediction quality | Hurt |
| Train on reconstruction within the conditional sequence rather than held-out future frames (ablation) | MSP · Sequential MNIST, learned representation | representation quality | Hurt |
| Use a learned nonlinear neural transition instead of a linear latent transition (ablation) | MSP · Sequential MNIST, extrapolation | prediction quality | Hurt |
| Extend the transition model to second order to represent acceleration (ablation) | Constant-velocity MSP · Accelerated Sequential MNIST | prediction quality | Improved |
| Apply constant-velocity prediction to sequences with constant acceleration (ablation) | MSP · Accelerated Sequential MNIST | prediction quality | Hurt |
| Apply MSP to rendered 3D rotations with potentially non-invertible transitions | MSP · Sequential ShapeNet | prediction quality | Hurt |
Closest papers by meaning
- Neural Fourier Transform: A General Approach to Equivariant Representation Learning (ICLR 2024 · similarity 0.86)
- Improving Equivariant Networks with Probabilistic Symmetry Breaking (ICLR 2025 · similarity 0.86)
- Learning Infinitesimal Generators of Continuous Symmetries from Data (NeurIPS 2024 · similarity 0.86)
- A Tale of Two Symmetries: Exploring the Loss Landscape of Equivariant Models (NeurIPS 2025 · similarity 0.85)
- MatrixNet: Learning over symmetry groups using learned group representations (NeurIPS 2024 · similarity 0.85)
- Probing Equivariance and Symmetry Breaking in Convolutional Networks (NeurIPS 2025 · 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.
