AI Paper Atlas

Research map / Optimization / Optimizers

SBO-RNN: Reformulating Recurrent Neural Networks via Stochastic Bilevel Optimization

Ziming Zhang, Yun Yue, Guojun Wu, Yanhua Li, Haichong Zhang

NeurIPS 2021 · Optimization › Optimizers · approach: Hypergradient bilevel optimization · 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
use dense U instead of sparse U=beta*I (SGD) (ablation)SBO-RNN · HAR-2 1% dataaccuracy: 79.18 -> 90.25Improved
use dense U instead of sparse U=beta*I (SGD) (ablation)SBO-RNN · HAR-2 10% dataaccuracy: 90.76 -> 93.25Improved
sparse U (1.5K vs 17.9K parameters) reduces training time (SGD) (ablation)SBO-RNN · HAR-2training time per batch: 0.60 -> 0.54Improved
replace SGD with heavy-ball momentum in lower level (dense) (ablation)SBO-RNN · HAR-2 1% dataaccuracy: 90.25 -> 86.59Hurt
replace SGD with Nesterov momentum in lower level (dense) (ablation)SBO-RNN · HAR-2 10% dataaccuracy: 93.25 -> 93.52No clear effect
replace SGD with momentum variants (ablation)SBO-RNN · HAR-2training time per batch: 0.60 -> 0.67Hurt
increase training data from 1% to 10% for sparse SGDSBO-RNN · HAR-2accuracy: 79.18 -> 90.76Improved
increase training data from 1% to 10% for dense SGDSBO-RNN · HAR-2accuracy: 90.25 -> 93.25Improved
increase hidden dimension from 128 to 256 (ablation)SBO-RNN · HAR-2 1% dataaccuracy: 83.99% -> 86.29%Improved
increase hidden dimension from 128 to 256 (ablation)SBO-RNN · HAR-2 full dataaccuracy: 95.08% -> 94.74%Hurt
increase hidden dimension from 128 to 256 (ablation)SBO-RNN · HAR-2 10% dataaccuracy: 91.31% -> 92.37%Improved
apply sparse-variant setup to SBO-RNN vs FastRNNSBO-RNN vs FastRNN · HAR-2 full dataaccuracy: 94.57% -> 94.96%Improved

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