Artificial Intelligence

Encoder-Decoder

An architecture where the encoder compresses input into a fixed representation and the decoder generates output from that representation. This structure is used in translation, summarization, and image captioning.

Why It Matters

The encoder-decoder paradigm is foundational — it is the basis for seq2seq models, transformers, and most modern generation systems.

Example

A machine translation system where the encoder processes an English sentence into a meaning vector and the decoder generates the equivalent French sentence from that vector.

Think of it like...

Like a telegraph system — one operator encodes a message into Morse code (encoder), sends it over the wire, and another operator decodes it back into words (decoder).

Related Terms