GTC 2020 S21252
Presenters: Vitaly Davydov,Poteha Labs
In most support systems, up to 70% of user questions are very similar to each other. Instead of manually answering each of them, it makes sense to automate it. We’ll discuss the business value of such systems, and how to integrate them into processes; how to develop an architecture for automating question-answering; how to set up a training loop, and which algorithms to choose; which models need to be trained, and why; and state-of-the-art language modeling models. Knowledge of Python, NLP, language models, and cloud computing will be helpful, but it’s not essential.
Watch this session
Join in the conversation below.