AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Efficient and Accurate Gradients for Neural SDEs

Patrick Kidger, James Foster, Xuechen Li, Terry Lyons

NeurIPS 2021 · Architectures and efficiency › RNNs and normalization · approach: Efficient network compression and training · 2 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
replace midpoint with reversible Heun (ablation)SDE-GAN · WeightsMMD (x1e-2): 4.38 -> 1.75Improved
replace midpoint with reversible Heun (ablation)SDE-GAN · Weightstraining time (days): 5.12 -> 2.59Improved
replace midpoint with reversible Heun (ablation)Latent SDE · Air qualitylabel classification accuracy (%): 46.3 -> 49.2Improved
replace midpoint with reversible Heun (ablation)Latent SDE · Air qualityMMD (x1e-2): 0.591 -> 0.472Improved
replace midpoint with reversible Heun (ablation)Latent SDE · Air qualitytraining time (hours): 5.58 -> 4.47Improved
replace Virtual Brownian Tree with Brownian Interval (ablation)SDE solver · SDE solve, size 32768time (s): 5.00 x 10^2 -> 4.7 x 10^1Improved
replace Virtual Brownian Tree with Brownian Interval (ablation)SDE solver · SDE solve, size 1time (s): 1.6 -> 8.2 x 10^-1Improved
replace Virtual Brownian Tree with Brownian Interval (ablation)SDE solver · doubly sequential access, size 32768time (s): 2.9 x 10^0 -> 3.5 x 10^-1Improved
replace gradient penalty with weight clipping (ablation)SDE-GAN with midpoint · OU datasetMMD (x1e-1): 2.58 -> 1.03Improved
replace gradient penalty with weight clipping (ablation)SDE-GAN with midpoint · OU datasettraining time (hours): 55.0 -> 32.5Improved
replace gradient penalty with weight clipping (ablation)SDE-GAN with midpoint · OU datasetreal/fake classification accuracy (%): 98.2 -> 93.9Improved
replace midpoint with reversible Heun under clipping (ablation)SDE-GAN · OU datasetMMD (x1e-1): 1.03 -> 0.45Improved

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