AI Paper Atlas

Research map / Optimization / Nonconvex and smooth optimization

Sensitivity-Free Gradient Descent Algorithms

Ion Matei, Maksym Zhenirovskyy, Johan de Kleer, John Maxwell

JMLR 2023 · Optimization › Nonconvex and smooth optimization · approach: Acceleration and Nesterov dynamics

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 sensitivity-enabled ODE solver with Algorithm 2 (no ODE solver, block coordinate descent)Algorithm 1 direct, Pytorch · Cucker-Smale N=5time per epoch (s): 0.463 -> 0.009 (50x)Improved
replace Algorithm 1 with Algorithm 2Algorithm 1 direct, Pytorch · Cucker-Smale N=200time per epoch (s): 0.614 -> 0.155 (4x)Improved
replace Algorithm 1 with Algorithm 3 (ODE solve, no sensitivities)Algorithm 1 direct, Pytorch · Cucker-Smale N=5time per epoch (s): 0.463 -> 0.171 (3x)Improved
use adjoint instead of direct sensitivity methodAlgorithm 1, Pytorch · Cucker-Smale N=200time per epoch (s): 0.614 -> 2.819Hurt
replace Algorithm 1 with Algorithm 2 in Jax with JITAlgorithm 1 adjoint, Jax · Cucker-Smale N=200time per epoch (s): 1.896 -> 0.0013 (x1458)Improved
replace Algorithm 1 with Algorithm 3 in JaxAlgorithm 1 adjoint, Jax · Cucker-Smale N=5time per epoch (s): 0.935 -> 0.06 (x16)Improved
replace Algorithm 1 with Algorithm 2Algorithm 1 direct, Pytorch · Cucker-Smale N=200 testRSSE on test data: 4.32 × 10−7 -> 3.24 × 10−8Improved
replace Algorithm 1 with Algorithm 2Algorithm 1, Jax · Cucker-Smale N=200 testRSSE on test data: 1.94 × 10−5 -> 1.46 × 10−7Improved
replace Algorithm 1 with Algorithm 3Algorithm 1, Jax · Cucker-Smale N=200 testRSSE on test data: 1.94 × 10−5 -> 2.17 × 10−5No clear effect
replace Algorithm 1 with Algorithm 2 for 4950 gamma parametersAlgorithm 1 direct · Cucker-Smale N=200, non-uniform gammaavg iteration time (s): 0.45 sec -> 0.06 secImproved
replace Algorithm 1 with Algorithm 2 for 4950 gamma parametersAlgorithm 1 direct · Cucker-Smale N=200, non-uniform gamma testRSSE: 3.5 × 10−4 -> 3.6 × 10−4Hurt
replace Algorithm 1 with Algorithm 3 for black-box modelAlgorithm 1 direct · Black-box neural ODE, Cucker-Smale N=200avg iteration time (s): 0.45 sec -> 0.29 secImproved

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