Enhancing Numerical Reasoning with the Guidance of Reliable Reasoning Processes

Numerical reasoning is an essential ability for NLP systems to handle numeric information. Recent research indicates that fine-tuning a small-scale model to learn generating reasoning processes alongside answers can significantly enhance performance. However, current methods have the limitation that most methods generate reasoning processes with large language models (LLMs), which are "unreliable" since such processes could contain information unrelated to the answer. To address this limitation, we introduce Enhancing NumeriCal reasOning with Reliable procEsses (ENCORE), which derives the reliable reasoning process by decomposing the answer formula, ensuring which fully supports the answer. Nevertheless, models could lack enough data to learn the reasoning process generation adequately, since our method generates only one single reasoning process for one formula. To overcome this difficulty, we present a series of pre-training tasks to help models learn the reasoning process generation with synthesized data. The experiments show that ENCORE yields improvement on all five experimental datasets with an average of 1.8%, proving the effectiveness of our method 1 . Currently, although LLMs have demonstrated 040 great performance on the numerical reasoning 041 (Chen et al., 2022a; Gao et al., 2022), we ar-042 gue that it is still valuable to study and employ 043 the small-scale model (e.g., BART LARGE (Lewis 044 et al., 2020)) since their low computational effi-045 ciency and decent performance, which still have ap-046 plication value in real scenarios. Previous research 047 has demonstrated that teaching small-scale models 048 to generate reasoning processes during fine-tuning 049 can make the prediction more accurate and explain-050