Yoshua Bengio
todayCanadian computer scientist whose work on language models and attention prefigured the transformer era. Founder of Mila, the largest deep-learning research institute in the world.
- Neural language models (2003).
- GANs co-credit and attention-mechanism work.
- Turing Award 2018; AI safety advocacy.
FAQ
What did Bengio's 2003 neural language model paper introduce?
The use of dense vector embeddings to represent words in a neural language model. The architecture predicted the next word from a fixed window of previous embeddings and, fifteen years later, became the template every transformer-based LLM still follows.
What is Mila?
The Quebec AI Institute, founded by Bengio in 1993 and now the largest academic deep-learning research group in the world. It hosts around a thousand researchers and is the centre of gravity for Canadian AI research.
What is Bengio's position on AI safety?
He has shifted publicly toward viewing frontier-AI risk as serious and near-term, has signed open letters calling for compute and capability governance, and now chairs the International AI Safety Report commissioned by 30 governments.
What was Bengio's role in attention and transformers?
His lab's 2014 paper with Bahdanau and Cho introduced attention to neural machine translation — the mechanism the 2017 transformer paper later generalised. Many researchers credit that 2014 work as the moment attention became central to modern NLP.