Qwen2.5-32B: Leveraging Self-Consistent Tool-Integrated Reasoning for Bengali Mathematical Olympiad Problem Solving

TL;DR

This paper introduces a novel deep learning approach utilizing the Qwen 2.5 series with Tool Integrated Reasoning and prompt engineering to effectively solve complex Bengali mathematical Olympiad problems, demonstrating the potential of NLP techniques in this domain.

Contribution

The paper presents a new methodology combining advanced deep learning models, prompt engineering, and Tool Integrated Reasoning to improve Bengali mathematical problem solving.

Findings

Enhanced problem-solving accuracy with TIR and prompt engineering.

Effective handling of complex calculations in Bengali math problems.

Robustness improved by removing RAG and tuning parameters.

Abstract

We present an innovative approach for solving mathematical problems in Bengali, developed for the DL Sprint 3.0 BUET CSE Fest 2024 Competition. Our method uses advanced deep learning models, notably the Qwen 2.5 series, with improvements made through prompt engineering, model quantization, and Tool Integrated Reasoning (TIR) to handle complex calculations. Initially, we explored various model architectures, including fine-tuned Mistral and quantized Qwen models, refining them with translation techniques, Retrieval-Augmented Generation (RAG), and custom dataset curation. Manual hyperparameter tuning optimized parameters like temperature and top-p to enhance model adaptability and accuracy. Removal of RAG and parameter adjustments further improved robustness. Our approach highlights the potential of advanced NLP techniques in solving Bengali mathematical problems.