New Yorker Caption Ranking: 250 Million Human Ratings on Cartoon Captions

Over 250 million human ratings on more than 2.2 million cartoon captions form a multimodal preference dataset for creative tasks. The dataset was created by researchers and is associated with a paper titled 'Humor in AI: Massive Scale Crowd-Sourced Preferences and Benchmarks for Cartoon Captioning'. It was last updated on the platform in September 2024.

Use Cases

• Benchmarking humor generation models based on the massive scale of crowd-sourced preferences.
• Training multimodal AI systems for creative captioning tasks based on cartoon-caption pairs.
• Studying human preferences and ranking algorithms for subjective, creative content.
• Developing models for caption ranking and selection based on the described preference data.

Strengths

• Contains over 250 million human ratings, indicating a very large scale of preference data.
• Covers more than 2.2 million cartoon captions, suggesting a broad corpus of creative content.
• Is explicitly designed for creative tasks and humor in AI, providing a focused research benchmark.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last updated 2024-09-15 04:09:32; freshness should be verified.

Provenance

Source

Crowd-sourced via the NextML caption contest data project.

Collection Method

Human ratings collected on cartoon captions.

Freshness

Last updated 2024-09-15 04:09:32.