Saliency Map
A visualisation technique that highlights the regions of an input, such as an image, that most strongly influence a model's prediction or classification.
In plain language
A highlighted overlay on an image showing what the AI looked at to make its decision. If an AI classifies a photo as a cat, the saliency map might light up around the ears and whiskers.
Why this matters
Saliency maps are relevant to implementation because they support model validation and debugging. They help developers understand whether models are using relevant features (ears and whiskers) or spurious patterns (background colours), contributing to more robust deployment decisions.
Relevance
ImplementationSaliency map analysis during model validation helps ensure that AI systems are basing decisions on features that are robust and interpretable to domain experts.
Related terms
Putting saliency map into practice in your organisation?
Ready to transform your AI strategy?
Partner with Australia's AI strategy and governance specialists. From adoption roadmaps to ISO 42001 audit readiness.