Research map / Architectures and efficiency / RNNs and normalization
Biologically-Plausible Learning Algorithms Can Scale to Large Datasets
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 sign-symmetric feedback weights instead of feedback alignment | ResNet-18 on ImageNet · ImageNet | performance | Helped |
| Preserve the error signal in the final layer while using feedback alignment or sign-symmetry in earlier layers (ablation) | Large-scale image classification models · ImageNet | performance | Helped |
| Use sign-symmetric feedback weights instead of backpropagation | ResNet-18 · ImageNet | performance | No clear effect |
| Replace sign-symmetric feedback weights with feedback alignment | Large-scale image classification models · ImageNet | performance | Hurt |
| Use sign-symmetric feedback weights | RetinaNet · MS COCO | object detection performance | Mixed |
| Multiply sign-symmetric feedback weights by a random matrix (ablation) | Sign-symmetry · ImageNet | performance | No clear effect |
| Reduce mini-batch size from 256 to 128 or 64 (ablation) | Sign-symmetry · ImageNet | initial-epoch performance | No clear effect |
| Add dropout to fc6 and fc7 in AlexNet or before the ResNet-18 fully connected layer (ablation) | Sign-symmetry · ImageNet | training speed | Hurt |
Closest papers by meaning
- Overcoming Rank Collapse in Feedback Alignment (NeurIPS 2026 · similarity 0.88)
- Efficient Target Propagation by Deriving Analytical Solution (AAAI 2024 · similarity 0.87)
- Improving equilibrium propagation without weight symmetry through Jacobian homeostasis (ICLR 2024 · similarity 0.86)
- Forward Learning with Top-Down Feedback: Empirical and Analytical Characterization (ICLR 2024 · similarity 0.86)
- One-Time Soft Alignment Enables Resilient Learning without Weight Transport (NeurIPS 2026 · similarity 0.85)
- Implicit Regularization in Feedback Alignment Learning Mechanisms for Neural Networks (ICML 2024 · 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.