Synthetic Data and Simulation

Generating synthetic data for training and testing, simulation and its environments, and the benefits and risks of this approach such as model collapse and the reproduction of bias. Excludes the use of simulation in learning world models and physical AI, which belongs to 3.7, privacy protection through anonymization, which belongs to 6.5, and documenting data provenance, which belongs to 6.4.

• 3.7 World Models and Physical AI
• 6.4 Quality, Documentation and Provenance
• 6.5 Privacy, Federated Learning and Anonymization

• 6.1 Data Lifecycle
• 6.2 Knowledge Graphs and Ontologies
• 6.4 Quality, Documentation and Provenance
• 6.5 Privacy, Federated Learning and Anonymization
• 6.6 Data Rights, Text and Data Mining, and Copyright
• 6.7 Interoperability and Shared Resources