Automatic End-to-End Data Integration using Large Language Models

Dimensions overall score 8.0

GitHub Code Pulse

No public code linked for this paper yet.

• An automatic data integration pipeline using GPT-5.2 can generate all necessary artifacts for specific use cases, including schema mappings, value mappings, training data for entity matching, and validation data for conflict resolution.method 70%
  Evidencepartial

  we present an automatic data integration pipeline that uses GPT-5.2 to generate all artifacts required to adapt the pipeline to specific use cases. These artifacts are schema mappings, value mappings for data normalization, training data for entity matching, and validation data for selecting conflict resolution heuristics in data fusion.

  Implicationpartial

  This is a core statement of the paper's contribution, explicitly detailed in the abstract.

  Verificationpartial

  partial

• The LLM-based data integration pipeline produces results comparable to, and in some cases better than, human-designed pipelines.result 70%
  Evidencepartial

  Our experiments show that the LLM-based pipeline is able to produce similar results, for some tasks even better results, as the human-designed pipelines.

  Implicationpartial

  The abstract directly compares the LLM pipeline's performance to human-designed pipelines and states it achieves similar or better results.

  Verificationpartial

  partial

• End-to-end, the LLM and human pipelines produce integrated datasets of comparable size and density.result 70%
  Evidencepartial

  End-to-end, the human and the LLM pipelines produce integrated datasets of comparable size and density.

  Implicationpartial

  This is a direct comparison of the output quality between the two pipeline types, stated explicitly in the abstract.

  Verificationpartial

  partial

• Configuring LLM-based data integration pipelines costs approximately $10 per case study.market 70%
  Evidencepartial

  Having the LLM configure the pipelines costs approximately $10 per case study

  Implicationpartial

  A specific cost figure is provided in the abstract for the LLM configuration.

  Verificationpartial

  partial

• The cost of LLM configuration for data integration pipelines is a small fraction of the cost of human data engineers performing the same tasks.market 70%
  Evidencepartial

  which represents only a small fraction of the cost of having human data engineers perform the same tasks.

  Implicationpartial

  The abstract directly contrasts the LLM configuration cost with human labor costs, highlighting the economic benefit.

  Verificationpartial

  partial

• Previous research has not investigated the potential of LLMs to replace all human input across end-to-end data integration pipelines.technical 70%
  Evidencepartial

  their potential to replace all human input across end-to-end data integration pipelines has not been investigated.

  Implicationpartial

  The abstract explicitly states this gap in prior research, framing the current work as novel.

  Verificationpartial

  partial

• Data integration pipelines typically require substantial manual effort from data engineers for configuration and training data labeling.technical 70%
  Evidencepartial

  Designing data integration pipelines typically requires substantial manual effort from data engineers to configure pipeline components and label training data.

  Implicationpartial

  This statement describes the traditional approach to data integration, providing context for the LLM solution.

  Verificationpartial

  partial

Startup potential card