Google Brain¶
Overview¶
Google Brain was Google's deep learning research division, founded in 2011. It merged with DeepMind in 2023 to form Google DeepMind. It was responsible for some of the most impactful research in modern AI, including the Transformer architecture.
Focus areas¶
- Deep learning research (neural architectures, training methods, optimisation)
- Large-scale distributed training (DistBelief, TensorFlow)
- Natural language processing and generation
- Reinforcement learning
- AI safety
Notable outputs (in this wiki)¶
- attention-is-all-you-need — "Attention Is All You Need" (Vaswani et al., 2017); introduced the Transformer.
- TensorFlow (open-source ML framework)
- BERT (2018), T5 (2019), PaLM (2022) — successor LLMs built on the Transformer
Connections¶
- attention-is-all-you-need — landmark paper from Google Brain.
- transformer-architecture — the Transformer is Google Brain's most consequential architectural contribution.