AI Paper Atlas

Research map / Learning theory and algorithms / Combinatorial and search optimization

Resilient Constrained Learning

Ignacio Hounie, Alejandro Ribeiro, Luiz F. O. Chamon

NeurIPS 2023 · Learning theory and algorithms › Combinatorial and search optimization · approach: Constrained neural network optimization · 1 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 resilient relaxation instead of hard constraintConstrained learning · Invariant MNIST, Rotated (90)test accuracy (%): 96.90 -> 97.19Improved
use resilient relaxation instead of hard constraintConstrained learning · Invariant MNIST, Scaledtest accuracy (%): 97.92 -> 98.20Improved
use resilient relaxation instead of hard constraintConstrained learning · Invariant MNIST, Translatedtest accuracy (%): 93.74 -> 95.21Improved
use resilient relaxation instead of hard constraintConstrained learning · Invariant MNIST, Rotated (180)test accuracy (%): 94.55 -> 95.38Improved
use resilient relaxation instead of hard constraintConstrained learning · Invariant MNIST, Originaltest accuracy (%): 98.74 -> 98.86Improved
compare resilient method with Augerino invariance learningResilient · Invariant MNIST, Rotated (180)test accuracy (%): 97.78 -> 95.38Hurt
compare resilient method with Augerino invariance learningResilient · Invariant MNIST, Scaledtest accuracy (%): 97.53 -> 98.20No clear effect
use resilient relaxation vs constrained baseline in federated learningConstrained FL · CIFAR10 imbalance 10clients improved (%): 77Mixed
use resilient relaxation vs constrained baseline in federated learningConstrained FL · F-MNIST imbalance 10max deterioration: 23.3Hurt
increase relaxation cost coefficient alpha (ablation)Resilient FL · Federated learning, class imbalancefinal training loss and perturbation normMixed
large alpha makes resilient problem approach constrained problem (ablation)Resilient FL · Federated learning, class imbalanceperturbation normImproved
use resilient instead of hard constraint on constraint feasibilityConstrained FL · Federated learning, class imbalancefraction infeasible clientsImproved

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