Open utterances-bot opened 1 year ago
This paper contributes on (1) simple human evaluation metric, (2) provide evidence correlation between perplexity and SSA, (3) end-to-end neural model with low perplexity has best performance.
https://psxxj-world.vercel.app/nlp-3
test
psxxj
commented
1 year ago
Towards a Human-like Open-Domain Chatbot
This paper contributes on (1) simple human evaluation metric, (2) provide evidence correlation between perplexity and SSA, (3) end-to-end neural model with low perplexity has best performance.
https://psxxj-world.vercel.app/nlp-3