We present Meena, a multi-turn open-domain chatbot trained end-to-end on data mined and filtered from public domain social media conversations. This 2.6B parameter neural network is simply trained to minimize perplexity of the next token. We also propose a human evaluation metric called Sensibleness and Specificity Average (SSA), which captures key elements of a human-like multi-turn conversation. Our experiments show strong correlation between perplexity and SSA. The fact that the best perplexity end-to-end trained Meena scores high on SSA (72% on multi-turn evaluation) suggests that a human-level SSA of 86% is potentially within reach if we can better optimize perplexity. Additionally, the full version of Meena (with a filtering mechanism and tuned decoding) scores 79% SSA, 23% higher in absolute SSA than the existing chatbots we evaluated.
Cite
@article{arxiv.2001.09977,
title = {Towards a Human-like Open-Domain Chatbot},
author = {Daniel Adiwardana and Minh-Thang Luong and David R. So and Jamie Hall and Noah Fiedel and Romal Thoppilan and Zi Yang and Apoorv Kulshreshtha and Gaurav Nemade and Yifeng Lu and Quoc V. Le},
journal= {arXiv preprint arXiv:2001.09977},
year = {2020}
}