Property-Driven Synthetic Data Engineering for Data-Scarce Software Systems: Reflections from the Breast Cancer Domain

2026-07-07 · Source: Artificial Intelligence · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Expert, quick

Summary

The paper introduces "property-driven synthetic data engineering" as a critical challenge in developing data-scarce software systems, particularly in sensitive domains like medicine. While synthetic data is often proposed for analysis, prediction, and testing in fields such as intraoperative radiotherapy (IORT) for breast cancer, the authors argue it merely shifts the engineering problem. The core issue becomes identifying, eliciting from stakeholders, validating under privacy constraints, and evolving the essential properties that synthetic data must preserve. Based on collaboration with oncologists and preliminary experiments with a sensitive IORT dataset, the authors highlight challenges across requirements, validation, privacy, and pipeline evolution. They advocate for automated software engineering research to develop methods and tools for formalizing, checking, and evolving these validity properties.

Key takeaway

For software engineers developing data-scarce systems in sensitive domains like healthcare, you must prioritize "property-driven synthetic data engineering." Focus on rigorously eliciting, formalizing, and validating the specific properties your synthetic data needs to preserve, rather than just generating data. This approach helps you manage privacy constraints and ensures the synthetic data remains relevant as system requirements evolve, mitigating risks associated with unvalidated data.

Key insights

Synthetic data shifts the engineering problem to defining and validating its required properties, especially in data-scarce domains.

Principles

• Synthetic data engineering requires property elicitation and validation.
• Privacy constraints complicate synthetic data validation.
• Validity properties for synthetic data must evolve.

Topics

• Synthetic Data Generation
• Data-Scarce Systems
• Software Engineering
• Breast Cancer Treatment
• Intraoperative Radiotherapy
• Data Privacy

Best for: AI Scientist, Research Scientist, Software Engineer

Related on AIssential

• Synthetic but Not Realistic: The Evaluation Challenge in Generative Modelling for Structured Electronic Medical Records — Current synthetic healthcare data evaluation methods overlook clinical validity, leading to unreliable inferences.
• A Pipeline for Generating Longitudinal Synthetic Clinical Notes Using Large Language Models — A modular LLM-powered pipeline generates privacy-preserving, longitudinal synthetic clinical notes for AI development and evaluation.
• Dozens of AI disease-prediction models were trained on dubious data — AI models for disease prediction are being trained on potentially fabricated public datasets, risking patient harm.
• From virtual experiments to biomedical insight with synthetic data — Bridging the "simulation to reality" gap is crucial for synthetic data to realize its full potential in biomedical AI.
• v333: Proceedings of CHIL 2026 — Machine learning, particularly LLMs, is rapidly expanding its utility across diverse clinical and administrative healthcare applications.
• A Pipeline for Generating Longitudinal Synthetic Clinical Notes Using Large Language Models — A modular LLM-powered pipeline generates consistent, diverse longitudinal synthetic clinical notes to enable privacy-preserving healthcare AI development.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.