Visual Question Answering (VQA)
Definition
Visual question answering provides linguistic answers to questions about an image.
A multimodal task where a system answers a natural language question about an image.
Simply explained
You show the AI an image and ask, for example: Which indicator is lit? It tries to find the answer in the image.
Why it matters in practice
The term helps to select, develop, evaluate AI systems in a well-founded manner and realistically assess their technical risks.
Example from business
A service team asks about visible indicators in a control cabinet photo but confirms safety-critical findings on site.
Limits and pitfalls
- Linguistic priors allow answers without image reference; benchmark score does not prove grounding or reliable perception.
Origin & history
Antol and coauthors published a seminal VQA dataset and benchmark in 2015.
Use in mid-sized companies
- Possible application: A service team asks about visible indicators in a control cabinet photo but confirms safety-critical findings on site.
Frequently asked questions
What does Visual Question Answering (VQA) mean?
A multimodal task where a system answers a natural language question about an image. What should you watch out for? Linguistic priors allow answers without image reference; benchmark score does not prove grounding or reliable perception.
Sources
- VQA · Antol et al., VQA
Source checked on 10.10.2026. Editorial application examples are hypothetical; no evidence of real customer results.
Editorially reviewed on: