Polite Dialogue Generation Without Parallel Data

By Tong Niu, Mohit Bansal. 2018

182 citations recorded by the source corpus.

Stylistic dialogue response generation, with valuable applications in personality-based conversational agents, is a challenging task because the response needs to be fluent, contextually-relevant, as well as paralinguistically accurate. Moreover, parallel datasets for regular-to-stylistic pairs are usually unavailable. We present three weakly-supervised models that can generate diverse, polite (or rude) dialogue responses without parallel data. Our late fusion model (Fusion) merges the decoder of an encoder-attention-decoder dialogue model with a language model trained on stand-alone polite utterances. Our label-finetuning (LFT) model prepends to each source sequence a politeness-score scaled label (predicted by our state-of-the-art politeness classifier) during training, and at test time is able to generate polite, neutral, and rude responses by simply scaling the label embedding by the corresponding score. Our reinforcement learning model (Polite-RL) encourages politeness generation by assigning rewards proportional to the politeness classifier score of the sampled response. We also present two retrievalbased, polite dialogue model baselines. Human evaluation validates that while the Fusion and the retrieval-based models achieve politeness with poorer context-relevance, the LFT and Polite-RL models can produce significantly more polite responses without sacrificing dialogue quality.

Subfields: Artificial Intelligence, Natural Language Processing, Generative Models

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

In Generative Models

Citations in the corpus

Recipes for Safety in Open-domain Chatbots (2020, background): “Niu and Bansal (2018) train a polite response generator that controls the degree of po-liteness of generations through scaling a control embedding according to a politeness score.”
Learning to summarize from human feedback (2020, background): “There is also a rich landscape of human feedback methods beyond binary comparisons that could be explored for training models [28, 17, 44, 64].”
Towards Empathetic Open-domain Conversation Models: A New Benchmark and Dataset (2018, methodology): “Similar methods have been used for controlling the style of generated text (e.g. Niu and Bansal, 2018).”
Politeness Transfer: A Tag and Generate Approach (2020, methodology): “Finally, we use a politeness classifier (Niu and Bansal, 2018) to assign politeness scores to these sentences and filter them into ten buckets based on the score (P0P9; Fig.”
PowerTransformer: Unsupervised Controllable Revision for Biased Language Correction (2020, background): “…to disentangle style from content and steer generations (e.g., Shen et al., 2017; Zhang et al., 2018; Fu et al., 2018; Yang et al., 2018; Niu and Bansal, 2018; Romanov et al., 2019; Dai et al., 2019; John et al., 2019) or backtranslation setups (Prabhumoye et al., 2018; Lample et al.,…”
Generating More Interesting Responses in Neural Conversation Models with Distributional Constraints (2018, background): “Niu and Bansal (2018) designed a number of weakly-supervised models that generate polite, neutral or rude responses.”
Can Neural Generators for Dialogue Learn Sentence Planning and Discourse Structuring? (2018, background): “Rao and Tetreault (2018) train a model to paraphrase from formal to informal style and Niu and Bansal (2018) use a high precision classiﬁer and a blended language model to control utterance politness.”
Controllable Story Generation with External Knowledge Using Large-Scale Language Models (2020, background): “…Generation Controllable text generation has a wide range of applications, including controlling through persona (Zhang et al., 2018; Boyd et al., 2020), politeness (Niu and Bansal, 2018), etc. Wiseman et al. (2018) presented controlling generations by learning latent, discrete templates from data.”
IMaT: Unsupervised Text Attribute Transfer via Iterative Matching and Translation (2019, background): “In the past few years, attribute-controlled NLG has witnessed renewed interest by researchers working on neural approaches to generation (Hu et al., 2017; Jhamtani et al., 2017; Melnyk et al., 2017; Mueller et al., 2017; Zhang et al., 2018; Prabhumoye et al., 2018; Niu and Bansal, 2018).”
Just Say No: Analyzing the Stance of Neural Dialogue Generation in Offensive Contexts (2021, background): “During training, these tokens are prepended to responses and at inference time, they are manually frozen to steer the model’s response towards the desired attribute (Niu and Bansal, 2018; See et al., 2019; Xu et al., 2020).”
Increasing Faithfulness in Knowledge-Grounded Dialogue with Controllable Features (2021, background): “Most relevant is the development of control-code-style input tokens such as in CTRL (Keskar et al., 2019) or the LFT model of Niu and Bansal (2018).”
Case Study: Deontological Ethics in NLP (2020, background): “, 1998), style (Niu and Bansal, 2018), content (Zhou et al.”
Diversifying Dialogue Generation with Non-Conversational Text (2020, methodology): “The idea has been widely utilized on domain adaptation for text generation tasks (Koehn and Schroeder, 2007; Wang et al., 2017; Niu and Bansal, 2018).”