A Neural Conversational Model

By Oriol Vinyals, Quoc V. Le. 2015

1,817 citations recorded by the source corpus.

Conversational modeling is an important task in natural language understanding and machine intelligence. Although previous approaches exist, they are often restricted to specific domains (e.g., booking an airline ticket) and require hand-crafted rules. In this paper, we present a simple approach for this task which uses the recently proposed sequence to sequence framework. Our model converses by predicting the next sentence given the previous sentence or sentences in a conversation. The strength of our model is that it can be trained end-to-end and thus requires much fewer hand-crafted rules. We find that this straightforward model can generate simple conversations given a large conversational training dataset. Our preliminary results suggest that, despite optimizing the wrong objective function, the model is able to converse well. It is able extract knowledge from both a domain specific dataset, and from a large, noisy, and general domain dataset of movie subtitles. On a domain-specific IT helpdesk dataset, the model can find a solution to a technical problem via conversations. On a noisy open-domain movie transcript dataset, the model can perform simple forms of common sense reasoning. As expected, we also find that the lack of consistency is a common failure mode of our model.

Subfields: Artificial Intelligence, Large Language Models, 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

In Large Language Models

In Natural Language Processing

Citations in the corpus

Deep Recurrent Models with Fast-Forward Connections for Neural Machine Translation (2016, background): “Moreover, NMT models can also be easily adapted to other tasks such as dialog systems (Vinyals and Le, 2015), question answering systems (Yu et al.”
Exploring the Limits of Language Modeling (2016, background): “…colors, facts about people, etc. Lastly, recently proposed sequence-to-sequence models employ conditional language models (Mikolov & Zweig, 2012) as their key component to solve diverse tasks like machine translation (Sutskever et al., 2014; Cho et al., 2014; Kalch-brenner et al., 2014) or video…”
Semi-supervised Sequence Learning (2015, methodology): “[32], which has been successfully used for machine translation [21, 11], text parsing [33], image captioning [35], video analysis [31], speech recognition [4] and conversational modeling [28, 34].”
MASS: Masked Sequence to Sequence Pre-training for Language Generation (2019, background): “…some inputs, including tasks like neural machine translation (NMT) (Cho et al., 2014; Bahdanau et al., 2015a; Vaswani et al., 2017), text summarization (Ayana et al., 2016; Suzuki & Nagata, 2017; Gehring et al., 2017) and conversational response generation (Shang et al., 2015; Vinyals & Le, 2015).”
CoQA: A Conversational Question Answering Challenge (2018, background): “Sequence-to-sequence (seq2seq) models have shown promising results for generating conversational responses (Vinyals and Le, 2015; Serban et al., 2016; Zhang et al., 2018).”
Sequence-Level Knowledge Distillation (2016, neutral): “…has been successfully applied to a wide range of tasks including parsing (Vinyals et al., 2015a), summarization (Rush et al., 2015), dialogue (Vinyals and Le, 2015; Serban et al., 2016; Li et al., 2016), NER/POS-tagging (Gillick et al., 2016), image captioning (Vinyals et al., 2015b; Xu et…”
The LAMBADA dataset: Word prediction requiring a broad discourse context (2016, background): “For example, the end-to-end system of Vinyals and Le (2015), trained on large conversational datasets, produces dialogues such as the following: (1) Human: what is your job?”
SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing (2018, background): “NMT has shown remarkable re-sults in several shared tasks (Denkowski and Neu-big, 2017; Nakazawa et al., 2017), and its effective approach has had a strong inﬂuence on other re-lated NLP tasks such as dialog generation (Vinyals and Le, 2015) and automatic summarization (Rush et al., 2015).”
Subword Regularization: Improving Neural Network Translation Models with Multiple Subword Candidates (2018, methodology): “Promising avenues for future work are to apply subword regularization to other NLP tasks based on encoder-decoder architectures, e.g., dialog generation (Vinyals and Le, 2015) and automatic summarization (Rush et al., 2015).”
Unsupervised Pretraining for Sequence to Sequence Learning (2016, background): “Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 383–391 Copenhagen, Denmark, September 7–11, 2017. c©2017 Association for Computational Linguistics”
LaMDA: Language Models for Dialog Applications (2022, background): “Dialog models [14, 15, 16], one of the most interesting applications of large language models, successfully take advantage of Transformers’ ability to represent long-term dependencies in text [17, 18].”
Towards a Human-like Open-Domain Chatbot (2020, background): “End-to-end neural network approaches (Shang et al., 2015; Vinyals and Le, 2015; Sordoni et al., 2015; Serban et al., 2016; Zhang et al., 2019), on the other hand, offer the simplicity of a single learned model.”
BLOOM: A 176B-Parameter Open-Access Multilingual Language Model (2022, background): “After Vinyals and Le (2015) observed limited task-performing behavior in a neural dialog model, Radford et al. (2019) later demonstrated that Transformer-based language models trained on text scraped from the web could perform various tasks to varying degrees.”
Anticipating Safety Issues in E2E Conversational AI: Framework and Tooling (2021, methodology): “These models are typically trained in the popular encoder-decoder paradigm, which was ﬁrst introduced for this task by Vinyals & Le (2015); Shang et al. (2015); Serban et al. (2016).”
Open-Domain Conversational Agents: Current Progress, Open Problems, and Future Directions (2020, background): “For example asking “what do you do for a living” and “what is your job?” a single agent would answer two different professions (Vinyals and Le, 2015), related to the consistency discussion above.”
Guiding the Release of Safer E2E Conversational AI through Value Sensitive Design (2022, methodology): “These models are typically trained in the popular encoder-decoder paradigm, which was first introduced for this task by Vinyals and Le (2015); Shang et al. (2015); Ser-ban et al. (2016).”
PALM: Pre-training an Autoencoding&autoregressive Language Model for Context-conditioned Generation (2020, background): “Conversational response generation aims to produce a ﬂexible response to a conversation (Vinyals and Le, 2015).”
ACUTE-EVAL: Improved Dialogue Evaluation with Optimized Questions and Multi-turn Comparisons (2019, methodology): “The two most used approaches, single-turn pairwise evaluation (Vinyals and Le, 2015; Li et al., 2016b), and multi-turn Likert scores (Venkatesh et al.”
Wizard of Wikipedia: Knowledge-Powered Conversational agents (2018, background): “The current state-of-the-art approaches, sequence to sequence models of various kinds (Sutskever et al., 2014; Vinyals & Le, 2015; Serban et al., 2016; Vaswani et al., 2017) attempt to address some of these skills, but generally suffer from an inability to bring memory and knowledge to bear; as…”