PAMELA: Human Aesthetic Ratings for AI-Generated Images

PAMELA is a dataset pairing AI-generated images with human ratings for aesthetic quality. The dataset includes rich participant demographics alongside each rating. It was created by researchers from the Tübingen AI Center and KU Leuven and was last updated on the Hugging Face platform in April 2026.

Use Cases

• Training aesthetic preference models based on human ratings.
• Analyzing demographic factors influencing image taste based on participant data.
• Benchmarking text-to-image generation systems for personalization.
• Studying the correlation between image features and human aesthetic judgments.

Strengths

• Pairs each image rating with rich participant demographics.
• Created by a multi-institution research team from Tübingen AI Center and KU Leuven.
• Last updated on 2026-04-10, indicating recent maintenance.

Limitations

• Row count, column definitions, and file formats are unknown, limiting suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect bias inherent to the specific human participant pool used for ratings.

Provenance

Source

bethgelab on Hugging Face.

Collection Method

AI-generated images rated by human participants.

Time Range

null

Freshness

Last updated 2026-04-10 14:52:10.

Geography

null