본문 바로가기
대학원 이야기/논문 리뷰

[논문 리뷰] You Impress Me: Dialogue Generation via Mutual Persona Perception (수정 중) (ACL 2020)

by misconstructed 2020. 11. 23.
728x90

 

논문 원본은 여기에서 확인할 수 있다.


 

 

 

 

 


# 참고 논문

Deep reinforcement learning for dialogue generation.

Personalizing dialogue agents: I have a dog, do you have pets too?

A persona-based neural conversation model. : speaker 에 대한 embedding 을 사용하는 방법 (리뷰)

Training millions of personalized dialogue agents.: PERSONA-CHAT dataset + Reddit dataset

TransferTransfo: A transfer learning approach for neural network based conversational agents. : GPT 를 기반으로 하는 방법 (리뷰)

Improving language understanding by generative pre-training. : GPT (리뷰)

Attention is all you need. : Transformer (리뷰)

Policy gradient methods for reinforcement learning with function approximation. : policy gradient methods for RL

Deal or no deal? endto-end learning of negotiation dialogues. : 협상을 위한 대화 학습, self-play 방식 사용

Dually interactive matching network for personalized response selection in retrieval-based chatbots. : retrieval-based 모델

Neural machine translation by jointly learning to align and translate. : Bahdanau attention (리뷰)

Sequence to sequence learning with neural networks. : seq2seq 모델 (리뷰)

Personalized response generation by dual-learning based domain adaptation. : dual-learning 기반

Aiming to know you better perhaps makes me a more engaging dialogue partner. : re-rank (beam search 기반)

An adversarial learning framework for a persona-based multi-turn dialogue model. : adversarial 기반

Personalizing dialogue agents via meta-learning. : meta-learning 기반

Listening between the lines: Learning personal attributes from conversations. : 평소 대화 속에서 사람의 특정을 찾아내는 방식 (리뷰)

Generating persona consistent dialogues by exploiting natural language inference. : persona 기반 response 생성 (1)

Exploiting persona information for diverse generation of conversational responses. : persona 기반 response 생성 (2)

Personalization in goal-oriented dialog. : goal-oriented 에서 personalization (1)

Learning personalized end-toend goal-oriented dialog.  : goal-oriented 에서 personalization (2)

728x90

댓글