MORE applies adaptive multi-objective RL to jointly optimize reasoning accuracy and linguistic quality in e-commerce dialogues, reporting 16.53% and 30.09% lifts in conversion metrics on ByteDance production traffic.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
One Model, Multiple Goals: Adaptive Multi-Objective Learning for E-commerce Dialogue Systems
MORE applies adaptive multi-objective RL to jointly optimize reasoning accuracy and linguistic quality in e-commerce dialogues, reporting 16.53% and 30.09% lifts in conversion metrics on ByteDance production traffic.