Pith. sign in

REVIEW

Adversarial Training Versus Weight Decay

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1804.03308 v3 pith:YSF2PAJ6 submitted 2018-04-10 cs.LG stat.ML

Adversarial Training Versus Weight Decay

classification cs.LG stat.ML
keywords trainingadversarialdecaymodelrobustnessweighteffectperturbations
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X LinkedIn Reddit HN
read the original abstract

Performance-critical machine learning models should be robust to input perturbations not seen during training. Adversarial training is a method for improving a model's robustness to some perturbations by including them in the training process, but this tends to exacerbate other vulnerabilities of the model. The adversarial training framework has the effect of translating the data with respect to the cost function, while weight decay has a scaling effect. Although weight decay could be considered a crude regularization technique, it appears superior to adversarial training as it remains stable over a broader range of regimes and reduces all generalization errors. Equipped with these abstractions, we provide key baseline results and methodology for characterizing robustness. The two approaches can be combined to yield one small model that demonstrates good robustness to several white-box attacks associated with different metrics.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.