Synthetic Data for Localized Solutions
summary
Synthetic data drives localized maternal‑health models for Kenya using simulation and GANs and the pipeline includes fidelity utility and privacy checks plus clinical‑utility evaluation. Practical details: 12,000 people per county recruited with medical tests and insurer partnerships to keep costs down, generators conditioned with metadata and gaussian noise for DP, and model types include cGAN, StyleGAN, timeGAN, CTGAN depending on modality. Deployed stack uses web and mobile with USSD fallback and datasets are shared within Kenyan healthcare partners.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:07Speaker background
- 2:13Synthetic data types
- 3:37Synthetic vs real data
- 6:03Localized solutions importance
- 8:31Cultural context factors
- 14:30Kenyan healthcare data
- 16:18Data collection challenges
- 19:50Synthetic data generation
- 21:55Impact on processes
- 22:17Key lessons
- 25:31Call to engage
- 26:32Q&A
- 27:07Unreliable real data
- 29:14Free-form text synthesis
- 31:20Sampling methods
- 32:30Concept drift detection
- 33:34Sharing datasets
related talks
30m🤯 No data? No problem! Synthetic data to the rescue
17mDon't Stop at Data Democratization!
20mSynthetic Data for Privacy, AI Fairness and Democratizing Data Access
36mGenerating realistic, differentially private data sets using GANs
25mThe Importance of Synthetic Data in Data-Centric AI
01h 00mThe Synthetic Data Advantage: How It Helps You to Get Both AI Privacy & Explainability
20mSynthetic Data: The new Super Power for AI
36mArchitecting Data: A Deep Dive Into the world of Synthetic Data