PLATO-2: Towards Building an Open-Domain Chatbot via Curriculum Learning

By Siqi Bao, Huang He, Fan Wang, Hua Wu, Haifeng Wang, Wenquan Wu, Zhen Guo, Zhibin Liu, Xinchao Xu. 2020

150 citations recorded by the source corpus.

To build a high-quality open-domain chatbot, we introduce the effective training process of PLATO-2 via curriculum learning. There are two stages involved in the learning process. In the first stage, a coarse-grained generation model is trained to learn response generation under the simplified framework of one-to-one mapping. In the second stage, a fine-grained generation model and an evaluation model are further trained to learn diverse response generation and response coherence estimation, respectively. PLATO-2 was trained on both Chinese and English data, whose effectiveness and superiority are verified through comprehensive evaluations, achieving new state-of-the-art results.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Large Language 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

1. Minlie Huang
2. Zhiyuan Liu
3. Yang Liu
4. Maosong Sun
5. Xipeng Qiu
6. Jianfeng Gao
7. Ji-rong Wen
8. Wayne Xin Zhao

In Computer Vision and Pattern Recognition

1. Qun Liu
2. Zhiyuan Liu
3. Minlie Huang
4. Xin Jiang
5. Yang Liu
6. Lifeng Shang
7. Radu Tudor Ionescu
8. Ji-rong Wen

In Large Language Models

1. Zhiyuan Liu
2. Qun Liu
3. Xin Jiang
4. Lifeng Shang
5. Jie Tang
6. Minlie Huang
7. Xu Han
8. Ji-rong Wen

Citations in the corpus

• Constructing Highly Inductive Contexts for Dialogue Safety through Controllable Reverse Generation (2022, methodology): “We augment 8 categories which have less than 3000 contexts using the responses sampled by the 3 dialogue models (Blender, DialoGPT and Plato2).”
• Pre-Trained Models: Past, Present and Future (2021, methodology): “These examples include Meena (Adiwardana et al., 2020), Blender (Roller et al., 2021), CDial-GPT (Wang et al., 2020e), Plato (Bao et al., 2020) and Plato-2 (Bao et al., 2021), which are trained on large-scale conversation data, commonly with the seq2seq framework.”
• DialGuide: Aligning Dialogue Model Behavior with Developer Guidelines (2022, unknown): “Current open-domain dialogue models such as Di-aloGPT (Zhang et al., 2020), Blenderbot (Roller et al., 2021), and PLATO (Bao et al., 2021) have shown the ability to generate fluent and interesting responses.”
• Empirical study on BlenderBot 2.0’s Errors Analysis in terms of Model, Data and User-Centric Approach (2022, unknown): “GPT-2 (Radford et al., 2019), Google’s Meena (Adiwardana et al., 2020), PLATO-2 (Bao et al., 2020), and BlenderBot 1.0 (Roller et al., 2020) all outperform existing encoder-decoder-based models.”
• Towards Identifying Social Bias in Dialog Systems: Framework, Dataset, and Benchmark (2022, background): “In recent years, significant efforts have been devoted to the development of open-domain dialog systems that are pre-trained on large-scale data to generate responses to user inputs (Freitas et al., 2020; Zhou et al., 2021a; Bao et al., 2021; Thop-pilan et al., 2022; Mi et al., 2022).”
• EVA: An Open-Domain Chinese Dialogue System with Large-Scale Generative Pre-Training (2021, methodology): “Our contributions are mainly as follows: (Wang et al., 2020) 12.0M 32.9M 380.1M 2.7 11.6 1.5GB PLATO-2 (Bao et al., 2020) 1.2B - - - - - STC (Shang et al., 2015) 4.4M 8.9M 158.1M 2 25.2 642MB Douban Conversation (Wu et al., 2017) 1.0M 7.1M 131.7M 6.7 18.6 535MB PersonalDialog (Zheng et al., 2019) 20”
• InstructDial: Improving Zero and Few-shot Generalization in Dialogue through Instruction Tuning (2022, methodology): “…dialogue domain, where models such as DialoGPT (Zhang et al., 2020), Meena (Adiwardana et al., 2020), Blenderbot (Roller et al., 2021) and PLATO (Bao et al., 2020, 2021) trained on sources such as Reddit, Twitter, Weibo, or on human-annotated datasets show great capabilities in carrying…”
• Dense Text Retrieval Based on Pretrained Language Models: A Survey (2022, background): “, Meena [338], BlenderBot [339], and PLATO-2 [340]) that pretrain dialog models on large-scale corpora have demonstrated their ability to converse like a human to some extent.”
• Preview, Attend and Review: Schema-Aware Curriculum Learning for Multi-Domain Dialogue State Tracking (2021, background): “…attracted increasing research interests in various NLP tasks, such as machine translation (Liu et al., 2020; Zhou et al., 2020), general language understanding (Xu et al., 2020), reading comprehension (Tay et al., 2019) and open-domain chatbots (Bao et al., 2020; Cai et al., 2020; Su et al., 2020).”
• Curriculum Learning: A Survey (2021, methodology): “[149] use a two-stage curriculum learning approach for building an open-domain chatbot.”
• On the Safety of Conversational Models: Taxonomy, Dataset, and Benchmark (2021, background): “Besides, Plato-2 is found weakest to control utterance-level safety.”