Introduces DDPO, a diversity-driven policy optimization algorithm, and a four-tier grading system to generate controllable, proficiency-matched spoken dialogues for K-12 English learners.
reason_consistency
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
-
Controllable Spoken Dialogue Generation: An LLM-Driven Grading System for K-12 Non-Native English Learners
Introduces DDPO, a diversity-driven policy optimization algorithm, and a four-tier grading system to generate controllable, proficiency-matched spoken dialogues for K-12 English learners.