This item was not updated in last three versions of the Radar. Should it have appeared in one of the more recent editions, there is a good chance it remains pertinent. However, if the item dates back further, its relevance may have diminished and our current evaluation could vary. Regrettably, our capacity to consistently revisit items from past Radar editions is limited.
Hold
Why?
Access to diverse and extensive datasets is crucial for training robust AI models, but sourcing this data can be challenging. Generative AI offers a solution to:
- Enhance dataset diversity, improving model performance on varied inputs.
- Increase the volume of training data, critical for deep learning models.
- Generate synthetic data where real data is scarce or sensitive.
What?
Utilizing generative AI techniques to:
- Create realistic, synthetic datasets that closely mimic genuine data attributes.
- Augment existing datasets, filling gaps and extending dataset utility without compromising privacy.
- Facilitate more efficient and effective model training across industries.
References
- Confluence page "Data Augmentation With Generative AI" in AD space
- "Data augmentation using Generative AI" section in the presentation from 2023-08-23
- Presentation "Synthetically generated datasets"
- Google Folder "Improving TKits on small datasets"