Domain Adaptive Dialog Generation via Meta Learning

By Kun Qian, Zhou Yu. 2019

134 citations recorded by the source corpus.

Domain adaptation is an essential task in dialog system building because there are so many new dialog tasks created for different needs every day. Collecting and annotating training data for these new tasks is costly since it involves real user interactions. We propose a domain adaptive dialog generation method based on meta-learning (DAML). DAML is an end-to-end trainable dialog system model that learns from multiple rich-resource tasks and then adapts to new domains with minimal training samples. We train a dialog system model using multiple rich-resource single-domain dialog data by applying the model-agnostic meta-learning algorithm to dialog domain. The model is capable of learning a competitive dialog system on a new domain with only a few training examples in an efficient manner. The two-step gradient updates in DAML enable the model to learn general features across multiple tasks. We evaluate our method on a simulated dialog dataset and achieve state-of-the-art performance, which is generalizable to new tasks.

Subfields: Artificial Intelligence, Natural Language Processing

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

1. Hinrich Schütze
2. Timo Schick
3. Minlie Huang
4. Andrea Madotto
5. Jianfeng Gao
6. Chenguang Zhu
7. Fei Mi
8. Zihan Liu

In Natural Language Processing

1. Hinrich Schütze
2. Timo Schick
3. Minlie Huang
4. Andrea Madotto
5. Zihan Liu
6. Fei Mi
7. Zi-Yi Dou
8. Baolin Peng

Citations in the corpus

• Exploiting Cloze-Questions for Few-Shot Text Classification and Natural Language Inference (2020, methodology): “Other approaches for few-shot learning in NLP include exploiting examples from related tasks (Yu et al., 2018; Gu et al., 2018; Dou et al., 2019; Qian and Yu, 2019; Yin et al., 2019) and using data augmentation (Xie et al.”
• Few-Shot Text Generation with Natural Language Instructions (2020, background): “…manually designed patterns and some amount of unlabeled examples, whereas meta learning approaches (e.g., Gu et al., 2018; Dou et al., 2019; Qian and Yu, 2019) require large annotated datasets for related tasks –, or they cannot be transferred to a generative setting in a straightforward…”
• Continual Learning for Natural Language Generation in Task-oriented Dialog Systems (2020, background): “The latter is a one-time adaptation process, and the focus is to optimize performance on a target domain transferred from source domains but without considering potential performance drop on source domains (Mi et al., 2019; Qian and Yu, 2019; Peng et al., 2020).”
• Meta Dialogue Policy Learning (2020, methodology): “The domain adaptive dialog generation method (DAML) [8] is an end-to-end dialogue system that can adapt to new domains with a few training samples.”
• MeLL: Large-scale Extensible User Intent Classification for Dialogue Systems with Meta Lifelong Learning (2021, background): “The applications of metalearning in NLP have not been frequently studied, with works such as [31, 32, 34].”
• Language Models as Few-Shot Learner for Task-Oriented Dialogue Systems (2020, background): “, 2019), pipelined endto-end models (Qian and Yu, 2019) and personalized dialogue systems (Madotto et al.”
• Investigating Meta-Learning Algorithms for Low-Resource Natural Language Understanding Tasks (2019, methodology): “Gu et al. (2018) have tried to apply first-order MAML in machine translation and Qian and Yu (2019) propose to address the domain adaptation problem in dialogue generation by using MAML.”
• Fast Domain Adaptation for Goal-Oriented Dialogue Using a Hybrid Generative-Retrieval Transformer (2020, background): “One promising direction that we are going to explore in our own future work is the meta-learning framework like used in [14].”
• Few-Shot Text Generation with Pattern-Exploiting Training (2020, background): “Other approaches for few-shot learning in NLP commonly require large sets of examples from re-lated tasks (Gu et al., 2018; Dou et al., 2019; Qian and Yu, 2019), parallel data for consistency training (Xie et al., 2019; Chen et al., 2020), or highly specialized methods tailored towards a specific…”
• Learning to Learn and Remember Super Long Multi-Domain Task Sequence (2022, background): “• For a personalized dialogue/recommendation system [44, 50], where learning the personal model for each user is viewed as an individual task, the user base may shift over time, e.g., the system is first deployed for Canadian users, then the company extends its market to Europe.”
• Recent advances in deep learning based dialogue systems: a systematic survey (2021, unknown): “Similarly, Qian and Yu (2019) also trained the end-to-end neural model using MAML to facilitate the domain adaptation, which enables the model to first train on rich-resource tasks and then on new tasks with limited data.”