AI Paper Atlas

Research map / Optimization / Optimizers

Full-Gradient Representation for Neural Network Visualization

Suraj Srinivas, François Fleuret

NeurIPS 2019 · Optimization › Optimizers · approach: Constrained and non-backprop training · 72 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
remove absolute-value post-processing from FullGrad (ablation)FullGrad · MNIST digit flipping (8 → 3)Δ log-odds: 8.26 ± 21.44 -> 12.93 ± 18.20Improved
remove absolute-value post-processing from FullGrad (ablation)FullGrad · MNIST pixel perturbation, RF = 0.5fractional output change: 0.06 ± 0.13 -> 0.19 ± 0.19Hurt
remove absolute-value post-processing from FullGrad (ablation)FullGrad · MNIST pixel perturbation, RF = 0.7fractional output change: 0.08 ± 0.11 -> 0.34 ± 0.23Hurt
remove absolute-value post-processing from FullGrad (ablation)FullGrad · MNIST pixel perturbation, RF = 0.9fractional output change: 0.55 ± 0.30 -> 0.63 ± 0.27Hurt
use FullGrad rather than IntegratedGradsaliency maps · MNIST pixel perturbation, RF = 0.5fractional output change: 0 ± 0 vs 0.06 ± 0.13Hurt
use FullGrad rather than IntegratedGradsaliency maps · MNIST pixel perturbation, RF = 0.7fractional output change: 0.004 ± 0.06 vs 0.08 ± 0.11Hurt
use FullGrad rather than IntegratedGradsaliency maps · MNIST pixel perturbation, RF = 0.9fractional output change: 0.44 ± 0.33 vs 0.55 ± 0.30Hurt
aggregate input-gradients with neuron bias-gradientsFullGrad · ImageNet 2012 validation pixel perturbationabsolute fractional output changeImproved
aggregate input-gradients with neuron bias-gradientsFullGrad · CIFAR100 remove-and-retrainaccuracy after retrainingImproved

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