Neural model seeks 'inappropriateness' to reduce chatbot awkwardness

Researchers from Skoltech and their colleagues from Mobile TeleSystems have introduced the notion of inappropriate text messages and released a neural model capable of detecting them, along with a large collection of such messages for further research. Among the potential applications are preventing corporate chatbots from embarrassing the companies that run them, forum post moderation, and parental control. The study came out in the Proceedings of the 8th Workshop on Balto-Slavic Natural Language Processing.

This post was originally published on this site

Lawyers Lookup - LawyersLookup.ca