Adding weight normalization to CrossQ stabilizes critic training and lets the algorithm scale to higher update-to-data ratios on continuous control benchmarks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Scaling CrossQ with Weight Normalization
Adding weight normalization to CrossQ stabilizes critic training and lets the algorithm scale to higher update-to-data ratios on continuous control benchmarks.