MTLinear clusters similarly correlated variates and scales losses by prediction error, making simple linear forecasters competitive with transformer-based state of the art on standard 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
-
A Multi-Task Learning Approach to Linear Multivariate Forecasting
MTLinear clusters similarly correlated variates and scales losses by prediction error, making simple linear forecasters competitive with transformer-based state of the art on standard benchmarks.