What is Cross-attention?
A mechanism where one sequence of data attends to another, enabling models to connect different inputs.
Why It Matters
Cross-attention is how translation models connect source and target languages, or how image captioning connects images to text.
Real-World Example
In translation, cross-attention links each word in the French output back to the relevant English input words.
“Understanding terms like Cross-attention matters because it helps you have better conversations with developers and make smarter decisions about your software. You do not need to be technical. You just need to know enough to ask the right questions.”
Related Terms
Self-attention
A mechanism where each word in a text considers its relationship to every other word in the same text.
Attention Mechanism
A technique that lets AI models focus on the most relevant parts of the input when generating output.
Encoder-Decoder
An AI architecture where one part compresses input into a representation and another part generates output from it.
Transformer
A type of AI architecture that processes text by paying attention to relationships between all words at once, rather than reading sequentially.
Learn More at buildDay Melbourne
Want to understand these concepts hands-on? Join our one-day workshop and build a real web application from scratch.
Related Terms
Transformer
A type of AI architecture that processes text by paying attention to relationships between all words at once, rather...
Attention Mechanism
A technique that lets AI models focus on the most relevant parts of the input when generating output.
Self-attention
A mechanism where each word in a text considers its relationship to every other word in the same text.
Encoder-Decoder
An AI architecture where one part compresses input into a representation and another part generates output from it.
Large Language Model (LLM)
An AI system trained on massive amounts of text that can understand and generate human language.
Tokenisation
The process of breaking text into smaller pieces called tokens that an AI model can process.