Semantically Aligned Question and Code Generation for Automated Insight Generation
This addresses a specific issue for knowledge workers like data scientists by improving the reliability of automated insights, though it is incremental as it builds on existing methods for alignment and filtering.
The paper tackles the problem of misalignment between generated insights and code in automated insight generation by using large language models to produce semantically aligned question-code pairs, and demonstrates through an empirical study on Open-WikiTable data that embeddings can filter out unaligned pairs and that joint generation increases question diversity.
Automated insight generation is a common tactic for helping knowledge workers, such as data scientists, to quickly understand the potential value of new and unfamiliar data. Unfortunately, automated insights produced by large-language models can generate code that does not correctly correspond (or align) to the insight. In this paper, we leverage the semantic knowledge of large language models to generate targeted and insightful questions about data and the corresponding code to answer those questions. Then through an empirical study on data from Open-WikiTable, we show that embeddings can be effectively used for filtering out semantically unaligned pairs of question and code. Additionally, we found that generating questions and code together yields more diverse questions.