tqa

北京航空航天大学大数据高精尖中心自然语言处理研究团队开展了智能问答的研究与应用总结。包括基于知识图谱的问答（KBQA），基于文本的问答系统（TextQA），基于表格的问答系统（TableQA）、基于视觉的问答系统（VisualQA）和机器阅读理解（MRC）等，每类任务分别对学术界和工业界进行了相关总结。

[⭐️EMNLP 2025 Findings]TabDSR: Decompose, Sanitize, and Reason for Complex Numerical Reasoning in Tabular Data

This application allows users to upload an .xlsx file, ask questions about the data within the file, and receive generated responses. The application leverages the Hugging Face transformers library to perform table question answering (TQA) and text generation to provide coherent, context-based answers.