Large Language Models in Research Automation

The field of research automation is witnessing significant advancements with the integration of large language models (LLMs). Recent developments indicate a shift towards leveraging LLMs for automating complex tasks such as workflow generation, schema discovery, and causal loop diagram creation. These innovations have the potential to transform the research landscape by improving efficiency, reducing manual effort, and enhancing the quality of research outputs. Noteworthy papers in this area include one that introduces a method for automating the translation of dynamic hypotheses into causal loop diagrams using LLMs with curated prompting techniques, achieving results comparable to expert-built diagrams. Another paper presents a framework for generating structured workflow outputs from sketch images using vision-language models, demonstrating improved performance over large vision-language models. Additionally, a survey highlights the potential of LLM-based scientific agents to automate critical tasks in scientific research, driving breakthroughs and advancing discovery.

Sources

Leveraging Large Language Models for Automated Causal Loop Diagram Generation: Enhancing System Dynamics Modeling through Curated Prompting Techniques

StarFlow: Generating Structured Workflow Outputs From Sketch Images

Scaling Laws of Scientific Discovery with AI and Robot Scientists

WorkTeam: Constructing Workflows from Natural Language with Multi-Agents

LLM-enabled Instance Model Generation

GIScience in the Era of Artificial Intelligence: A Research Agenda Towards Autonomous GIS

DebFlow: Automating Agent Creation via Agent Debate

SchemaAgent: A Multi-Agents Framework for Generating Relational Database Schema

Towards Scientific Intelligence: A Survey of LLM-based Scientific Agents

Generating Structured Plan Representation of Procedures with LLMs

Leveraging LLMs for User Stories in AI Systems: UStAI Dataset

LLMs4SchemaDiscovery: A Human-in-the-Loop Workflow for Scientific Schema Mining with Large Language Models

Geospatial Artificial Intelligence for Satellite-based Flood Extent Mapping: Concepts, Advances, and Future Perspectives

Built with on top of