ML: Compression

Machine-learning ideas tagged Compression in the ML taxonomy of the Math2NN corpus.

2 ideas found

✗ Failed on benchmark 2026

Quadratic center-manifold bottleneck

Replace a conventional linear decoder in an autoencoder or latent state-space model with an explicit quadratic manifold decoder, allowing a small latent vector to represent curved and transport-like state trajectories. Add a dynamics-aware invariance loss that penalizes the discrepancy between the time derivative of the quadratic manifold and the neural dynamics evaluated on that manifold.

Useful7/10

Difficulty5/10

Novelty6/10

Architecture Accuracy Memory Sample efficiency

Paper: Beyond linear subspaces: Nonlinear moment matching meets quadratic manifolds arXiv:2608.19486

Unverified 2026

Entropy-Neutral Stream Mixing for Rate-Aware Networks

Place doubly stochastic stream mixing immediately before a quantizer, activation compressor, or latent bottleneck and jointly optimize task loss with estimated code length. The paper's entropy argument says that this linear mixing cannot increase differential entropy, so it can provide cross-stream representation capacity without an ideal entropy-rate penalty; the entropy bottleneck then learns which feature values deserve bits.

Useful6/10

Difficulty4/10

Novelty4/10

Memory Memory Scalability Accuracy

Paper: Resource-efficient Semantic Coding Schemes with Manifold-constrained Hyper-connections arXiv:2608.13253