AssistedDS: Benchmarking How External Domain Knowledge Assists LLMs in Automated Data Science

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Luo, An, Xian, Xun, Du, Jin, Tian, Fangqiao, Wang, Ganghua, Zhong, Ming, Zhao, Shengchun, Bi, Xuan, Liu, Zirui, Zhou, Jiawei, Srinivasa, Jayanth, Kundu, Ashish, Fleming, Charles, Hong, Mingyi, Ding, Jie
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917035062067200
author Luo, An
Xian, Xun
Du, Jin
Tian, Fangqiao
Wang, Ganghua
Zhong, Ming
Zhao, Shengchun
Bi, Xuan
Liu, Zirui
Zhou, Jiawei
Srinivasa, Jayanth
Kundu, Ashish
Fleming, Charles
Hong, Mingyi
Ding, Jie
author_facet Luo, An
Xian, Xun
Du, Jin
Tian, Fangqiao
Wang, Ganghua
Zhong, Ming
Zhao, Shengchun
Bi, Xuan
Liu, Zirui
Zhou, Jiawei
Srinivasa, Jayanth
Kundu, Ashish
Fleming, Charles
Hong, Mingyi
Ding, Jie
contents Large language models (LLMs) have advanced the automation of data science workflows. Yet it remains unclear whether they can critically leverage external domain knowledge as human data scientists do in practice. To answer this question, we introduce AssistedDS (Assisted Data Science), a benchmark designed to systematically evaluate how LLMs handle domain knowledge in tabular prediction tasks. AssistedDS features both synthetic datasets with explicitly known generative mechanisms and real-world Kaggle competitions, each accompanied by curated bundles of helpful and adversarial documents. These documents provide domain-specific insights into data cleaning, feature engineering, and model selection. We assess state-of-the-art LLMs on their ability to discern and apply beneficial versus harmful domain knowledge, evaluating submission validity, information recall, and predictive performance. Our results demonstrate three key findings: (1) LLMs frequently exhibit an uncritical adoption of provided information, significantly impairing their predictive performance when adversarial content is introduced, (2) helpful guidance is often insufficient to counteract the negative influence of adversarial information, and (3) in Kaggle datasets, LLMs often make errors in handling time-series data, applying consistent feature engineering across different folds, and interpreting categorical variables correctly. These findings highlight a substantial gap in current models' ability to critically evaluate and leverage expert knowledge, underscoring an essential research direction for developing more robust, knowledge-aware automated data science systems. Our data and code are publicly available here: https://github.com/jeremyxianx/Assisted-DS
format Preprint
id arxiv_https___arxiv_org_abs_2506_13992
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AssistedDS: Benchmarking How External Domain Knowledge Assists LLMs in Automated Data Science
Luo, An
Xian, Xun
Du, Jin
Tian, Fangqiao
Wang, Ganghua
Zhong, Ming
Zhao, Shengchun
Bi, Xuan
Liu, Zirui
Zhou, Jiawei
Srinivasa, Jayanth
Kundu, Ashish
Fleming, Charles
Hong, Mingyi
Ding, Jie
Machine Learning
Artificial Intelligence
Computation and Language
Methodology
62-07, 62-08, 68T05, 68T07, 68T01, 68T50
I.2.0; I.2.6; I.2.7; I.5.1; I.5.4; H.2.8; G.3
Large language models (LLMs) have advanced the automation of data science workflows. Yet it remains unclear whether they can critically leverage external domain knowledge as human data scientists do in practice. To answer this question, we introduce AssistedDS (Assisted Data Science), a benchmark designed to systematically evaluate how LLMs handle domain knowledge in tabular prediction tasks. AssistedDS features both synthetic datasets with explicitly known generative mechanisms and real-world Kaggle competitions, each accompanied by curated bundles of helpful and adversarial documents. These documents provide domain-specific insights into data cleaning, feature engineering, and model selection. We assess state-of-the-art LLMs on their ability to discern and apply beneficial versus harmful domain knowledge, evaluating submission validity, information recall, and predictive performance. Our results demonstrate three key findings: (1) LLMs frequently exhibit an uncritical adoption of provided information, significantly impairing their predictive performance when adversarial content is introduced, (2) helpful guidance is often insufficient to counteract the negative influence of adversarial information, and (3) in Kaggle datasets, LLMs often make errors in handling time-series data, applying consistent feature engineering across different folds, and interpreting categorical variables correctly. These findings highlight a substantial gap in current models' ability to critically evaluate and leverage expert knowledge, underscoring an essential research direction for developing more robust, knowledge-aware automated data science systems. Our data and code are publicly available here: https://github.com/jeremyxianx/Assisted-DS
title AssistedDS: Benchmarking How External Domain Knowledge Assists LLMs in Automated Data Science
topic Machine Learning
Artificial Intelligence
Computation and Language
Methodology
62-07, 62-08, 68T05, 68T07, 68T01, 68T50
I.2.0; I.2.6; I.2.7; I.5.1; I.5.4; H.2.8; G.3
url https://arxiv.org/abs/2506.13992