Model Adaptation · Software component
Synthetic Data Generator
Software componentModel AdaptationModelsarc:SyntheticDataGenerator
A model-adaptation component that prompts a generative model to produce many synthetic task trajectories or example variations from human seed examples or tutorial-derived task goals.
Responsibility. Scales training and demonstration data by LLM-based generation.
Also known as: LLM-based trajectory generation, Trajectory synthesis, Synthetic demonstration generation, Synthetic data augmentation stage, Fairness data augmenter, Counterfactual data generator
Relationships
invokes dependency
- LLM Inference Service Ch3.5 Ch7.5
- LLM Judge abstract Ch3.5
is invoked by dependency
writes dependency
receives data from dynamic
is orchestrated by control
produces lifecycle
Design guidance
- SHOULD start from human seed examples and have human reviewers validate a random sample of generations.
- SHOULD NOT be used as the sole data source without filtering, since fine-tuned models inherit the generator's biases and failure modes.
- SHOULD generate synthetic data only for underrepresented categories where real examples are unavailable.
- SHOULD keep synthetic data to at most 5-10% of total training data to avoid teaching LLM stylistic artifacts.
Quantitative guidance
As stated by the sources; verify before use.
- Can generate ~10,000 trajectories overnight for API cost versus tens of thousands of expert hours manually (Ch3.5).
- Example: 10,000 synthetic ESG analysis documents provide baseline coverage for a zero-coverage category (Ch7.5).
Classification
- Patterns
- Hybrid human-seed plus LLM-scale generationTutorial-guided trajectory synthesisParaphrasing / back-translation augmentationCounterfactual data generation
- Technologies
- AgentTrekNVIDIA NeMo data curationGPT-4NVIDIA NeMo Curator
- Quality attributes
- Performance efficiency (ISO/IEC 25010)Cost efficiency
- Risks mitigated
- Data scarcity in specialised domainsProhibitive expert annotation costCoverage gaps in low-resource subdomains after aggressive filtering
Sources
- Ch3.5: T. Nguyen, "Prompt Optimization, Few-Shot Learning, Fine-Tuning," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 3.5. ISBN: 9798244538229.
- Ch7.5: T. Nguyen, "NeMo Curator, Riva Speech AI & Multimodal Integration," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 7.5. ISBN: 9798244538229.
- Ch9.4: T. Nguyen, "Fairness and Bias Mitigation," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 9.4. ISBN: 9798244538229.