A two-step framework combining block-wise imputation with shared and task-specific multi-task learning improves prediction under distribution and posterior shift.
In all numerical tasks, we set the decay constant to 0 .95, applied every 200 iterations
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
-
Multi-task Learning for Heterogeneous Multi-source Block-Wise Missing Data
A two-step framework combining block-wise imputation with shared and task-specific multi-task learning improves prediction under distribution and posterior shift.