Research map / Optimization / Implicit bias and generalization
On the Theory of Implicit Deep Learning: Global Convergence with Implicit Layers
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 |
|---|---|---|---|
| use a deep equilibrium linear model instead of a linear model | Modified Kuzushiji-MNIST · Modified Kuzushiji-MNIST | test error | Helped |
| use a deep equilibrium linear model instead of a DNN with ReLU | Modified Kuzushiji-MNIST · Modified Kuzushiji-MNIST | test error | Helped |
| use a deep equilibrium linear model instead of a linear model | Modified CIFAR-10 · Modified CIFAR-10 | test error | Helped |
| use a deep equilibrium linear model instead of a DNN with ReLU | Modified CIFAR-10 · Modified CIFAR-10 | test error | Helped |
| use a deep equilibrium linear model instead of a linear model | Modified CIFAR-100 · Modified CIFAR-100 | test error | Helped |
| use a deep equilibrium linear model instead of a DNN with ReLU | Modified CIFAR-100 · Modified CIFAR-100 | test error | Helped |
| use a deep equilibrium linear model instead of a linear ResNet with identity initialization | Deep equilibrium linear model · Gaussian synthetic data | training loss | Helped |
| use a deep equilibrium linear model instead of a linear ResNet with identity initialization | Deep equilibrium linear model · Uniform synthetic data | training loss | No clear effect |
Closest papers by meaning
- A global convergence theory for deep ReLU implicit networks via over-parameterization (ICLR 2022 · similarity 0.90)
- Robust Implicit Networks via Non-Euclidean Contractions (NeurIPS 2021 · similarity 0.90)
- DeltaDEQ: Exploiting Heterogeneous Convergence for Accelerating Deep Equilibrium Iterations (NeurIPS 2024 · similarity 0.89)
- Subhomogeneous Deep Equilibrium Models (ICML 2024 · similarity 0.89)
- Separation and Bias of Deep Equilibrium Models on Expressivity and Learning Dynamics (NeurIPS 2024 · similarity 0.88)
- Positive Concave Deep Equilibrium Models (ICML 2024 · similarity 0.87)
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.