Word Problem

Research on solving and generating word problems, particularly in mathematics and physics, focuses on leveraging large language models (LLMs) to improve accuracy and interpretability. Current efforts involve developing methods to enhance LLMs' reasoning capabilities, often through symbolic representations of problems and the incorporation of domain-specific knowledge, such as ontologies, to ensure consistency and accuracy in generated problems. These advancements aim to automate the creation of educational materials and improve the accessibility of problem-solving tools, ultimately impacting both educational practices and the development of more robust AI systems.