Daniil Burlakov, Head of the Recommendations Group, Plus and FinTech, Yandex.
Speaker: holds a Candidate of Sciences degree (PhD equivalent) in physics and mathematics; after postgraduate studies he joined Yandex's School of Data Analysis, then Yandex itself, and has worked on recommendations since 2016. The talk covers how the recommendation system into which the transformer is embedded works, the evolution of the algorithms and the product, the transformer's architecture and training, implementation results, and lessons learned.
Continue reading — it's free
Google or a one-time email code. No password.
