Skip to content

Latest commit

 

History

History
3 lines (2 loc) · 516 Bytes

README.md

File metadata and controls

3 lines (2 loc) · 516 Bytes

Dialogue-Agent

We explore the PersonaChat dataset in the context of both zero-shot and few-shot conditions, addressing the former by extracting explicit personas from dialogue history and the latter by using model-agnostic meta-learning as well as contrastive learning to learn with a few dialogues with personas and then perform the bulk of the learning on datasets without such personas. We also explore improvements yielded by augmenting the dialogue history with external information relevant to the dialogue