Reproduce “Scalable Equilibrium Propagation via Intermediate Error Signals for Deep Convolutional CRNNs” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jiaqi Lin, Malyaban Bal, Abhronil Sengupta
The authors add auxiliary learning signals at intermediate layers to equilibrium propagation, helping deeper convolutional recurrent networks train when signals fade through the network. Local-error and knowledge-distillation signals guide selected layers alongside the output loss. On CIFAR-10, VGG-7-KD reports 89.67% test accuracy versus 88.60% for VGG-5 EP; architectures differ, and evaluation is limited to CIFAR VGG networks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.