Search papers, labs, and topics across Lattice.
The AgentDS benchmark was introduced to evaluate AI agents and human-AI collaboration on 17 domain-specific data science challenges across six industries. Results from a competition with 29 teams revealed that AI-only agents perform near or below the median of human participants, indicating a struggle with domain-specific reasoning. The strongest solutions arose from human-AI collaboration, highlighting the continued importance of human expertise.
Despite advances in LLMs, human-AI collaboration still significantly outperforms AI-only agents in domain-specific data science tasks, proving that human expertise remains crucial.
Data science plays a critical role in transforming complex data into actionable insights across numerous domains. Recent developments in large language models (LLMs) and artificial intelligence (AI) agents have significantly automated data science workflow. However, it remains unclear to what extent AI agents can match the performance of human experts on domain-specific data science tasks, and in which aspects human expertise continues to provide advantages. We introduce AgentDS, a benchmark and competition designed to evaluate both AI agents and human-AI collaboration performance in domain-specific data science. AgentDS consists of 17 challenges across six industries: commerce, food production, healthcare, insurance, manufacturing, and retail banking. We conducted an open competition involving 29 teams and 80 participants, enabling systematic comparison between human-AI collaborative approaches and AI-only baselines. Our results show that current AI agents struggle with domain-specific reasoning. AI-only baselines perform near or below the median of competition participants, while the strongest solutions arise from human-AI collaboration. These findings challenge the narrative of complete automation by AI and underscore the enduring importance of human expertise in data science, while illuminating directions for the next generation of AI. Visit the AgentDS website here: https://agentds.org/ and open source datasets here: https://huggingface.co/datasets/lainmn/AgentDS .
