DeepAnalyze: Agentic Large Language Models for Autonomous Data Science

2025-10-19

AI Summary

Existing data science agents rely on predefined workflows, limiting their adaptability to open-ended, complex tasks. Method: We propose the first autonomous data science agent based on a large language model (LLM), introducing a curriculum-style agent training paradigm and a data-anchored trajectory synthesis framework. Using an 8B-parameter model, we perform end-to-end training via agent-centric reinforcement learning augmented with high-quality synthetic trajectories. Contribution/Results: Our agent breaks free from rigid workflow constraints, enabling full-stack autonomous reasoning and execution—from raw data ingestion to analyst-level, in-depth reporting. Remarkably, despite its modest 8B parameter count, it outperforms state-of-the-art closed-source workflow-based agents. To foster community advancement, we publicly release the model, training code, and synthetic trajectory dataset—significantly accelerating the development of autonomous data science as a research paradigm.

Technology Category

• Multiagent Systems: Modeling other Agents
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Autonomous data science, from raw data sources to analyst-grade deep research reports, has been a long-standing challenge, and is now becoming feasible with the emergence of powerful large language models (LLMs). Recent workflow-based data agents have shown promising results on specific data tasks but remain fundamentally limited in achieving fully autonomous data science due to their reliance on predefined workflows. In this paper, we introduce DeepAnalyze-8B, the first agentic LLM designed for autonomous data science, capable of automatically completing the end-toend pipeline from data sources to analyst-grade deep research reports. To tackle high-complexity data science tasks, we propose a curriculum-based agentic training paradigm that emulates the learning trajectory of human data scientists, enabling LLMs to progressively acquire and integrate multiple capabilities in real-world environments. We also introduce a data-grounded trajectory synthesis framework that constructs high-quality training data. Through agentic training, DeepAnalyze learns to perform a broad spectrum of data tasks, ranging from data question answering and specialized analytical tasks to open-ended data research. Experiments demonstrate that, with only 8B parameters, DeepAnalyze outperforms previous workflow-based agents built on most advanced proprietary LLMs. The model, code, and training data of DeepAnalyze are open-sourced, paving the way toward autonomous data science.

Problem

• Achieving fully autonomous data science without predefined workflows
• Automating end-to-end pipeline from raw data to research reports
• Enabling LLMs to progressively acquire complex data science capabilities

Innovation

• Agentic LLM for autonomous data science pipeline
• Curriculum-based training emulating human learning trajectory
• Data-grounded framework synthesizing high-quality training trajectories