Google created the artificial-intelligence architecture called Transformer in 2017.
The Transformer was introduced in the research paper “Attention Is All You Need,” written by researchers associated with Google Brain and Google Research. Its central innovation was the use of attention mechanisms to process relationships between words or other elements in a sequence.
Unlike many earlier sequence models, Transformers could process sequence elements more efficiently in parallel during training. This design made them highly effective for language translation and later for large language models, image systems, and other generative-AI applications.
The architecture is often confused with a single commercial chatbot or model. Transformer is a general neural-network architecture. Google’s research introduced it, while many other organizations later built systems based on the same underlying design.