TL;DR

The Transformer replaces recurrence and convolutions entirely with stacked self-attention layers, enabling full sequence parallelization and setting new state-of-the-art on machine translation.

Summary

Sequence-to-sequence models relying on RNNs suffer from sequential computation that blocks parallelization and degrades on long-range dependencies. The Transformer architecture addresses this with multi-head self-attention: queries, keys, and values are projected into multiple subspaces and attended in parallel. Positional encodings (sinusoidal) preserve sequence order without recurrent state. An encoder-decoder structure with cross-attention links source and target. Trained on WMT14, it reached 28.4 BLEU (EN-DE) and 41.0 BLEU (EN-FR) at a fraction of prior compute cost.

Key contributions

  1. Introduces the Transformer — a sequence model built purely from attention and feedforward layers, with no recurrence or convolutions.
  2. Proposes multi-head attention as the core building block, attending to multiple representation subspaces simultaneously.
  3. Establishes sinusoidal positional encoding as a parameter-free way to inject sequence order into attention-based models.
  4. Demonstrates that full parallelization over sequence positions reduces training time by orders of magnitude versus RNN baselines.
  5. Shows state-of-the-art BLEU scores on WMT14 EN-DE and EN-FR with lower computational cost than all prior models.

When to cite

  1. When explaining the architectural foundation of any modern LLM (GPT, BERT, T5, and all descendants).
  2. When arguing why self-attention is preferable to RNNs for capturing long-range dependencies.
  3. When discussing training efficiency and why large-scale pretraining became feasible.
  4. When introducing multi-head attention or positional encoding as concepts in a technical explanation.

LLM AI Generative AI Deep Learning