Deep learning has achieved breakthroughs in natural language processing and generation tasks. In this talk we will look at the potential and the limitations of seq2seq models for building open-domain chatbots with human-like traits, showcasing our most recent project, "a mean, grumpy, sarcastic chatbot in the browser". The project was accepted as artwork in the online gallery at the NeurIPS 2018 Workshop Machine Learning for Creativity and Design.

Suzana Ilic will present this talk. She is based in Tokyo and currently contracting with Google Japan as a Data Project Manager in Machine Intelligence: Natural Language Understanding.

She is leading Machine Learning Tokyo, a community-based group of ML engineers and researchers with bi-weekly deep learning meetups and workshops and more than 1000 members.

About the event

This event will include drinks and snacks.

Please reserve your space below, numbers are limited. There will be an entry fee of $50 at the door (the fee is waived for Ooosh Members).