Common-Mode Collapse and Recovery in Direct Feedback Alignment
cs.LG, cs.NE
Submitted: 2026-09-25
Updated: 2026-09-25
Code: https://github.com/KempnerInstitute/DFA-Stall
Terminology
Sources
- Overcoming Rank Collapse in Feedback Alignment
- The Influence of Learning Rule on Representation Dynamics in Wide Neural Networks
- Gaussian Error Linear Units (GELUs)
- Understanding Dimensional Collapse in Contrastive Self-supervised Learning
- Adam: A Method for Stochastic Optimization
- Principled Training of Neural Networks with Direct Feedback Alignment
- Direct Feedback Alignment Provides Learning in Deep Neural Networks
- Conditioned Direct Feedback Alignment via Activity and Error Geometry
- Benchmarking the Accuracy and Robustness of Feedback Alignment Algorithms
- Exact solutions to the nonlinear dynamics of learning in deep linear neural networks
- Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks