Research map / Probabilistic and scientific modeling / Dynamical systems
Improved Expressivity Through Dendritic Neural Networks
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 DENN dendritic layer instead of standard FNN architectures | FNN · 121 UCI classification datasets | average rank: 3.08 vs 4.05-4.86 (vs SNN 3.50, p=1.37e-01) | Helped |
| DENN vs self-normalizing network (SNN) | DENN · 121 UCI classification datasets | average rank: 3.08 vs 3.50 | No clear effect |
| use DENN with mid-size branch numbers instead of standard FNN | FNN · Fashion-MNIST and CIFAR-10 training loss | training loss | Helped |
| use extreme branch numbers (2 or 256) in DENN (ablation) | DENN · Fashion-MNIST and CIFAR-10 training loss | training loss | No clear effect |
| use DENN under capacity pressure (small layer sizes) | FNN · CIFAR-100 training accuracy and loss | training accuracy | Helped |
| replace Maxout nonlinearity in DENN with ReLU-based branch integration (ablation) | DENN · permutation-invariant image datasets | fitting performance | Hurt |
| use Maxout network with more kernels at constant synapse count | Maxout · Fashion-MNIST and CIFAR-10 training loss | training loss | Hurt |
| train on random Gaussian noise data instead of image data (ablation) | DENN · random data fitting | training loss advantage over FNN | Hurt |
| use DENN on image datasets for test generalization | FNN · image datasets test accuracy | test accuracy | No clear effect |
Closest papers by meaning
- Dendritic Localized Learning: Toward Biologically Plausible Algorithm (ICML 2025 · similarity 0.88)
- The Expressive Leaky Memory Neuron: an Efficient and Expressive Phenomenological Neuron Model Can Solve Long-Horizon Tasks. (ICLR 2024 · similarity 0.87)
- Dendritic Integration Inspired Artificial Neural Networks Capture Data Correlation (NeurIPS 2024 · similarity 0.86)
- Scaling Laws and Tradeoffs in Recurrent Networks of Expressive Neurons (NeurIPS 2026 · similarity 0.85)
- Shaping the distribution of neural responses with interneurons in a recurrent circuit model (NeurIPS 2024 · similarity 0.84)
- Normative Networks for Source Separation via Local Plasticity and Dendritic Computation (NeurIPS 2026 · similarity 0.84)
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.