Knowledge in Triples for LLMs: Enhancing Table QA Accuracy with Semantic Extraction

Integrating structured knowledge from tabular formats poses significant challenges within natural language processing (NLP), mainly when dealing with complex, semi-structured tables like those found in the FeTaQA dataset. These tables require advanced methods to interpret and generate meaningful responses accurately. Traditional approaches, such as SQL and SPARQL, often fail to fully capture the semantics of such data, especially in the presence of irregular table structures like web tables. This paper addresses these challenges by proposing a novel approach that extracts triples straightforward from tabular data and integrates it with a retrieval-augmented generation (RAG) model to enhance the accuracy, coherence, and contextual richness of responses generated by a fine-tuned GPT-3.5-turbo-0125 model. Our approach significantly outperforms existing baselines on the FeTaQA dataset, particularly excelling in Sacre-BLEU and ROUGE metrics. It effectively generates contextually accurate and detailed long-form answers from tables, showcasing its strength in complex data interpretation.

Large-Scale Data Mining of Rapid Residue Detection Assay Data From HTML and Documents: Improving Data Access and Visualization for Veterinarians

Extra-label drug use in food animal medicine is authorized by the US Animal Medicinal Drug Use Clarification Act (AMDUCA), and estimated withdrawal intervals are based on published scientific pharmacokinetic data. Occasionally there is a paucity of scientific data on which to base a withdrawal interval or a large number of animals being treated, driving the need to test for drug residues. Rapid assay commercial farm-side tests are essential for monitoring drug residues in animal products to protect human health. Active ingredients, sensitivity, matrices, and species that have been evaluated for commercial rapid assay tests are typically reported on manufacturers' websites or in PDF documents that are available to consumers but may require a special access request. Additionally, this information is not always correlated with FDA-approved tolerances. Furthermore, parameter changes for these tests can be very challenging to regularly identify, especially those listed on websites or in documents that are not publicly available. Therefore, artificial intelligence plays a critical role in efficiently extracting the data and ensure current information. Extracting tables from PDF and HTML documents has been investigated both by academia and commercial tool builders. Research in text mining of such documents has become a widespread yet challenging arena in implementing natural language programming. However, techniques of extracting tables are still in their infancy and being investigated and improved by researchers. In this study, we developed and evaluated a data-mining method for automatically extracting rapid assay data from electronic documents. Our automatic electronic data extraction method includes a software package module, a developed pattern recognition tool, and a data mining engine. Assay details were provided by several commercial entities that produce these rapid drug residue assay