Corpus ID: 210920238

Towards a Human-like Open-Domain Chatbot

@article{Adiwardana2020TowardsAH,
  title={Towards a Human-like Open-Domain Chatbot},
  author={Daniel De Freitas 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},
  year={2020},
  volume={abs/2001.09977}
}
  • Daniel De Freitas Adiwardana, Minh-Thang Luong, +8 authors Quoc V. Le
  • Published 2020
  • Computer Science, Mathematics
  • ArXiv
  • 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… CONTINUE READING

    Citations

    Publications citing this paper.
    SHOWING 1-10 OF 32 CITATIONS

    Recipes for building an open-domain chatbot

    VIEW 16 EXCERPTS
    CITES BACKGROUND, METHODS & RESULTS
    HIGHLY INFLUENCED

    Large Scale Multi-Actor Generative Dialog Modeling

    VIEW 2 EXCERPTS
    CITES BACKGROUND
    HIGHLY INFLUENCED

    FILTER CITATIONS BY YEAR

    2019
    2020

    CITATION STATISTICS

    • 6 Highly Influenced Citations

    • Averaged 17 Citations per year from 2019 through 2020

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 63 REFERENCES

    A Neural Conversational Model

    VIEW 1 EXCERPT

    Unifying Human and Statistical Evaluation for Natural Language Generation

    VIEW 5 EXCERPTS
    HIGHLY INFLUENTIAL