Neural Networks
Encoder-Decoder
Neural Networks· Intermediate
Definition
A neural architecture with two components: an encoder that compresses input into a latent representation, and a decoder that generates output from that representation. Foundational to machine translation, summarization, and many seq2seq tasks. Contrasts with decoder-only architectures (GPT family) dominant in current LLMs.
Enterprise Context
Understanding encoder-decoder vs. decoder-only tradeoffs helps organizations select the right model class: encoder-decoders excel at structured transformation tasks; decoder-only at open-ended generation.
Tags
#architecture#seq2seq#transformer
MS
Maxx Stacks Editorial
Reviewed by enterprise AI practitioners
Maxx University
Keep learning. Keep building.
250+ terms. 5 learning paths. AI maturity assessment. Jargon translator. All free, always.