Proposes backward-coherence regularization turning RNN hidden-state sequences into quasi-reverse-martingales, yielding almost-sure convergence and empirical gains on ICU, economic, and activity data.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
Develops target-oriented statistical compression where conditional target processes form reverse martingales, with defects measuring loss in approximate summaries, applied to sequential boundary monitoring.
citing papers explorer
-
Target-Oriented Statistical Compression: Sufficiency, Reverse Martingales, and Sequential Monitoring
Develops target-oriented statistical compression where conditional target processes form reverse martingales, with defects measuring loss in approximate summaries, applied to sequential boundary monitoring.