Research map / Architectures and efficiency / RNNs and normalization
Neural Path Features and Neural Path Kernel : Understanding the role of gates in deep learning
Chandrashekar Lakshminarayanan, Amit Vikram Singh
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 |
|---|---|---|---|
| learn gates in decoupled feature network (DLNPF) instead of fixed random gates (GCONV, Adam) (ablation) | deep gated network · CIFAR-10 | test accuracy (%): 67.09 -> 77.12 | Improved |
| learn gates in DLNPF instead of fixed random gates (FC, Adam) (ablation) | deep gated network · MNIST | test accuracy (%): 96.02 -> 98.22 | Improved |
| use fixed learnt gates (FLNPF) instead of fixed random gates (GCONV, Adam) (ablation) | deep gated network · CIFAR-10 | test accuracy (%): 67.09 -> 79.68 | Improved |
| use fixed learnt gates (FLNPF) instead of fixed random gates (VCONV, SGD) (ablation) | deep gated network · CIFAR-10 | test accuracy (%): 58.92 -> 63.06 | Improved |
| make feature and value network initialisation independent (FRNPF-II) versus identical (FRNPF-DI) (ablation) | fixed random NPF · MNIST and CIFAR-10 | test accuracy (%): 67.09 -> 67.08 | No clear effect |
| replace ReLU network with FLNPF (GCONV, Adam) (ablation) | ReLU CNN · CIFAR-10 | test accuracy (%): 80.32 -> 79.68 | No clear effect |
| use gates from later training stage (FLNPF-1 to FLNPF-6) (ablation) | FLNPF · CIFAR-10 GCONV | test accuracy (%): 72 -> 79.43 | Improved |
| learn gates (FLNPF-6) versus infinite width CNTK | GCONV · CIFAR-10 | test accuracy (%): 77.43 -> 79.43 | Improved |
| use finite width with fixed random gates versus CNTK | GCONV FRNPF · CIFAR-10 | test accuracy (%): 77.43 -> 67.09 | Hurt |
| use Adam instead of SGD for GCONV DLNPF | deep gated network · CIFAR-10 | test accuracy (%): 74.57 -> 77.12 | Improved |
| add global average pooling (GCONV versus VCONV) with ReLU, Adam | ReLU CNN · CIFAR-10 | test accuracy (%): 72.43 -> 80.32 | Improved |
Closest papers by meaning
- ReLUs Are Sufficient for Learning Implicit Neural Representations (ICML 2024 · similarity 0.82)
- The Effect of Weight Precision on the Neuron Count in Deep ReLU Networks (ICML 2024 · similarity 0.80)
- Gompertz Linear Units: Leveraging Asymmetry for Enhanced Learning Dynamics (NeurIPS 2025 · similarity 0.80)
- Activation-Descent Regularization for Input Optimization of ReLU Networks (ICML 2024 · similarity 0.79)
- Compelling ReLU Networks to Exhibit Exponentially Many Linear Regions at Initialization and During Training (ICML 2025 · similarity 0.79)
- Make Haste Slowly: A Theory of Emergent Structured Mixed Selectivity in Feature Learning ReLU Networks (ICLR 2025 · similarity 0.78)
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.
