Dynamic linear and exponential schedules for softness parameters in soft-hard BPGD improve rate-distortion performance and reduce non-convergence for LDGM-based lossy source coding while avoiding grid searches.
Sufficient conditions for c onvergence of the Sum-Product Algorithm
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IT 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Dynamic Parameter Scheduling in Soft-Hard BPGD for Lossy Source Coding
Dynamic linear and exponential schedules for softness parameters in soft-hard BPGD improve rate-distortion performance and reduce non-convergence for LDGM-based lossy source coding while avoiding grid searches.