A distributional RL framework with information bottleneck achieves 37-41% better DRAM equalizer performance than baselines, with 51x speedup and uncertainty quantification via Monte Carlo dropout and CVaR.
Deep reinforcement learning-based optimal and fast hybrid equalizer design method for high-bandwidth memory (hbm) module,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Distributional Reinforcement Learning with Information Bottleneck for Uncertainty-Aware DRAM Equalization
A distributional RL framework with information bottleneck achieves 37-41% better DRAM equalizer performance than baselines, with 51x speedup and uncertainty quantification via Monte Carlo dropout and CVaR.