Tool-Augmented Hybrid Ensemble Reasoning with Distillation for Bilingual Mathematical Problem Solving
Peiqing Lu, Yuan Zhang, Haoyun Zhang, Jiasen Zheng, Kejian Tong, Wenjun Wu · 2025
Read Tool-Augmented Hybrid Ensemble Reasoning with Distillation for Bilingual Mathematical Problem Solving by Peiqing Lu, Yuan Zhang, Haoyun Zhang, Jiasen Zheng, Kejian Tong, Wenjun Wu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-19093.txt.
Description
Peiqing Lu, Yuan Zhang, Haoyun Zhang, and Jiasen Zheng built HERALD, a system for solving math problems written in two languages. Language models read problems well but often slip on exact calculation. HERALD joins several models, including NuminaMath-7B-TIR, GPT-4o, and Mistral-7B, with external math tools.
Adaptive routing sends each problem down a suitable reasoning path, and tool-based reinforcement learning decides when to call a calculator or symbolic tool. Knowledge distillation then compresses the ensemble so answers come faster. Confidence checks and dual-path verification guard the final result. The authors report gains in accuracy, stability, and clarity for bilingual problem solving.
Catalog record
| CREATOR | Peiqing Lu, Yuan Zhang, Haoyun Zhang, Jiasen Zheng, Kejian Tong, Wenjun Wu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-19093