IRIS interleaves staged curriculum supervised fine-tuning with reverse-curriculum reinforcement learning using a composite reward to improve mathematical reasoning in English and low-resource Indian languages, accompanied by a new 29k-problem multilingual dataset.
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
-
IRIS: Interleaved Reinforcement with Incremental Staged Curriculum for Cross-Lingual Mathematical Reasoning
IRIS interleaves staged curriculum supervised fine-tuning with reverse-curriculum reinforcement learning using a composite reward to improve mathematical reasoning in English and low-resource Indian languages, accompanied by a new 29k-problem multilingual dataset.