Research map / Architectures and efficiency / RNNs and normalization
Deep Predictive Coding Network for Object Recognition
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 |
|---|---|---|---|
| add recursive feedforward and feedback processing | Plain-A on CIFAR-10 · CIFAR-10 test | accuracy (%): 90.61 -> 93.83 | Helped |
| add recursive feedforward and feedback processing | Plain-A on CIFAR-100 · CIFAR-100 test | accuracy (%): 62.11 -> 72.58 | Helped |
| add recursive feedforward and feedback processing | Plain-D on SVHN · SVHN test | error rate (%): 3.21(3.41±0.13) -> 2.28(2.42±0.09) | Helped |
| add recursive feedforward and feedback processing | Plain-E on SVHN · SVHN test | error rate (%): 3.19(3.41±0.13) -> 2.24(2.42±0.10) | Helped |
| add recursive feedforward and feedback processing | Plain-D on MNIST · MNIST test | error rate (%): 0.53(0.59±0.04) -> 0.38(0.46±0.06) | Helped |
| add recursive feedforward and feedback processing | Plain-E on MNIST · MNIST test | error rate (%): 0.68(0.74±0.03) -> 0.36(0.48±0.06) | Helped |
| share feedforward and feedback weights (ablation) | PCN-A-6 on CIFAR-10 · CIFAR-10 test | accuracy (%): 93.83 -> 92.26 | Hurt |
| share feedforward and feedback weights (ablation) | PCN-A-6 on CIFAR-100 · CIFAR-100 test | accuracy (%): 72.58 -> 69.44 | Hurt |
| increase recursive computation cycles (ablation) | PCN · CIFAR-10 and CIFAR-100 test | accuracy (%) | Helped |
| use PCN rather than its feedforward-only counterpart | Matched PCN and plain CNN architectures · CIFAR-10/100, SVHN, and MNIST | classification performance | Helped |
Closest papers by meaning
- Benchmarking Predictive Coding Networks -- Made Simple (ICLR 2025 · similarity 0.90)
- Training a Predictive Coding Network on ImageNet using Equilibrium Propagation (NeurIPS 2026 · similarity 0.88)
- A Stable, Fast, and Fully Automatic Learning Algorithm for Predictive Coding Networks (ICLR 2024 · similarity 0.87)
- $\mu$PC: Scaling Predictive Coding to 100+ Layer Networks (NeurIPS 2025 · similarity 0.87)
- Only Strict Saddles in the Energy Landscape of Predictive Coding Networks? (NeurIPS 2024 · similarity 0.86)
- Auto-Compressing Networks (NeurIPS 2025 · similarity 0.85)
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.