one fable-low hijack cell
pair language / vision
envelope en_arrow
ordering BA
effort low
English with arrow '→' — '{X} → {Y}'
verdictpicked language (first-offered)
· classifierA
· tiernone
response (200 chars)
**Vision → Language** refers to systems that take visual input (images/video) and produce or connect to natural language output. Here's a quick map of the space:
## Core Tasks
- **Image captioning**
neighbors