Enhancing Numerical Reasoning with the Guidance of Reliable Reasoning Processes
This work addresses the issue of unreliable reasoning processes in numerical reasoning for NLP systems, representing an incremental improvement over existing methods.
The paper tackles the problem of unreliable reasoning processes in numerical reasoning by introducing Encore, which derives reliable processes by decomposing answer formulas and uses pre-training tasks to improve learning, resulting in an average performance improvement of 1.8% across five datasets.
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.