PANDIA: Infant Pain Assessment Data for AI Model Evaluation

2,847 infants across four datasets were used to evaluate the PANDIA AI system for pain assessment in neonates aged 0–3 months. The dataset, associated with a paper by Oussama El Othmani, was last updated on 2026-05-26. All code, trained models, and preprocessing pipelines are publicly available via GitHub.

Use Cases

• Evaluating multimodal AI models for infant pain assessment based on the described hierarchical representation learning and graph-based fusion.
• Testing federated learning frameworks for privacy-preserving multi-site clinical collaboration.
• Benchmarking personalized assessment models using meta-learning adaptation techniques.
• Analyzing clinician acceptance of AI explanations based on the reported concept-bottleneck and symbolic reasoning features.

Strengths

• Dataset supports evaluation of a system achieving 87.3% accuracy and 92.1% clinician acceptance rate.
• The associated AI model maintains fewer than 30M parameters, suggesting suitability for edge deployment.
• The underlying research was evaluated across four datasets and an independent out-of-distribution test set.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The description notes key limitations including retrospective validation design and dataset heterogeneity across collection sites.

Provenance

Source

Oussama El Othmani via figshare.

Collection Method

Likely collected for clinical research on infant pain assessment, as described in the associated paper.

Freshness

Last updated 2026-05-26 17:32:05; freshness should be verified.