Research map / Learning theory and algorithms / Sample complexity
Gaussian Lower Bound for the Information Bottleneck Limit
Amichai Painsky, Naftali Tishby
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 |
|---|---|---|---|
| replace naive separate Gaussianization with mixture-aware custom Gaussianization | separate Gaussianization lower bound · Univariate Gaussian mixture example | Gaussian lower bound on MI (bits): 0.0628 -> 0.204 | Improved |
| use AGCE instead of the custom mixture-aware Gaussianization | Gaussian lower bound on MI · Univariate Gaussian mixture example | Gaussian lower bound on MI (bits): 0.204 -> 0.411 | Improved |
| use separate Gaussianization of ACE output instead of AGCE | Gaussian lower bound on MI · Univariate Gaussian mixture example | Gaussian lower bound on MI (bits): 0.411 -> 0.389 | No clear effect |
| replace separate Gaussianization with bi-terminal Gaussianization of ACE output | multivariate Gaussian lower bound · Multivariate toy models, d>5 with KCCA | Gaussian lower bound on MI | Improved |
| use KCCA instead of ACE for dimension above 5 | multivariate Gaussian lower bound · Multivariate toy models | Gaussian lower bound on MI | Hurt |
| use the Gaussian lower bound instead of GIB applied directly to non-monotonically scrambled data | IB curve lower bound · Exponential model, IB curve | IB curve lower bound | Improved |
Closest papers by meaning
- Information Bottleneck Analysis of Deep Neural Networks via Lossy Compression (ICLR 2024 · similarity 0.86)
- Cauchy-Schwarz Divergence Information Bottleneck for Regression (ICLR 2024 · similarity 0.85)
- Structured IB: Improving Information Bottleneck with Structured Feature Learning (AAAI 2025 · similarity 0.83)
- Partial Information Decomposition via Normalizing Flows in Latent Gaussian Distributions (NeurIPS 2025 · similarity 0.83)
- The Breakdown of Gaussian Universality in Classification of High-dimensional Linear Factor Mixtures (ICLR 2025 · similarity 0.83)
- Neural Mutual Information Estimation with Vector Copulas (NeurIPS 2025 · similarity 0.82)
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.
