AI Paper Atlas

Research map / Architectures and efficiency / Spiking networks

Direct Training of SNN using Local Zeroth Order Method

Bhaskar Mukhoty, Velibor Bojkovic, William de Vazelhes, Xiaohan Zhao, Giulia De Masi, Huan Xiong, Bin Gu

NeurIPS 2023 · Architectures and efficiency › Spiking networks · approach: Training deep spiking networks · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (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 changedWhereMeasuredEffect
add LocalZO to TETTET (ResNet-19) · CIFAR10accuracy: 94.50 -> 95.56Improved
add LocalZO to TETTET (ResNet-19) · CIFAR100accuracy: 74.72 -> 77.25Improved
add LocalZO to tDBNtDBN (ResNet-19) · CIFAR10accuracy: 93.16 -> 95.07Improved
add LocalZO to tDBNtDBN (ResNet-19) · CIFAR100accuracy: 71.12 -> 73.74Improved
add LocalZO to TETTET (VGGSNN) · DVS-CIFAR10accuracy: 74.89 -> 75.62Improved
add LocalZO to tDBN vs TET on neuromorphic dataLocalZO (VGGSNN) · N-Caltech-101accuracy: 74.65 -> 79.86Improved
replace surrogate with LocalZO in SparseGrad frameworkSparseGrad · NMNIST (Normal, m=1)test accuracy: 91.86 -> 93.29Improved
replace surrogate with LocalZO in SparseGrad frameworkSparseGrad · FMNIST (Normal, m=1)test accuracy: 80.28 -> 81.79Improved
replace surrogate with LocalZO in SparseGrad frameworkSparseGrad · SHD (Normal, m=1)test accuracy: 74.73 -> 76.55Improved
replace SparseGrad with LocalZO (speedup cost)SparseGrad · NMNIST (Normal, m=1)overall speedup: 3.38 -> 3.34No clear effect
increase number of samples m from 1 to 100 (ablation)LocalZO · NMNISToverall speedup: 3.33 -> 1.59Hurt
increase number of samples m from 1 to 100 (ablation)LocalZO · NMNISTtest accuracy: 93.29 -> 93.81Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy