InsightEval: An Expert-Curated Benchmark for Assessing Insight Discovery in LLM-Driven Data Agents
This work addresses the problem of evaluating insight discovery in data analysis for researchers using LLMs and multi-agent systems, but it is incremental as it builds upon and improves existing benchmarks.
The authors tackled the lack of benchmarks for evaluating insight discovery capabilities in LLM-driven data agents by identifying flaws in existing frameworks like InsightBench and proposing essential criteria for a high-quality benchmark. They developed InsightEval, a new dataset with a novel metric, and used it to highlight challenges and guide future research in automated insight discovery.
Data analysis has become an indispensable part of scientific research. To discover the latent knowledge and insights hidden within massive datasets, we need to perform deep exploratory analysis to realize their full value. With the advent of large language models (LLMs) and multi-agent systems, more and more researchers are making use of these technologies for insight discovery. However, there are few benchmarks for evaluating insight discovery capabilities. As one of the most comprehensive existing frameworks, InsightBench also suffers from many critical flaws: format inconsistencies, poorly conceived objectives, and redundant insights. These issues may significantly affect the quality of data and the evaluation of agents. To address these issues, we thoroughly investigate shortcomings in InsightBench and propose essential criteria for a high-quality insight benchmark. Regarding this, we develop a data-curation pipeline to construct a new dataset named InsightEval. We further introduce a novel metric to measure the exploratory performance of agents. Through extensive experiments on InsightEval, we highlight prevailing challenges in automated insight discovery and raise some key findings to guide future research in this promising direction.