Smart Reply: Automated Response Suggestion for Email

Abstract

In this paper we propose and investigate a novel end-to-end method for automatically generating short email responses, called Smart Reply. It generates semantically diverse suggestions that can be used as complete email responses with just one tap on mobile. The system is currently used in Inbox by Gmail and is responsible for assisting with 10% of all mobile responses. It is designed to work at very high throughput and process hundreds of millions of messages daily. The system exploits state-of-the-art, large-scale deep learning. We describe the architecture of the system as well as the challenges that we faced while building it, like response diversity and scalability. We also introduce a new method for semantic clustering of user-generated content that requires only a modest amount of explicitly labeled data.

Accepted to KDD 2016

References in corpus (2)

• Sequence to Sequence Learning with Neural Networks
• Neural Responding Machine for Short-Text Conversation

Cited by in corpus (36)

• Fundamentals of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) Network
• Scaling Graph Neural Networks with Approximate PageRank
• Efficient Natural Language Response Suggestion for Smart Reply
• Embedding Projector: Interactive Visualization and Interpretation of Embeddings
• LaMPost: Design and Evaluation of an AI-assisted Email Writing Prototype for Adults with Dyslexia
• Generative Deep Neural Networks for Dialogue: A Short Review
• Adversarial Evaluation of Dialogue Models
• Training Neural Response Selection for Task-Oriented Dialogue Systems
• ProjectionNet: Learning Efficient On-Device Deep Networks Using Neural Projections
• Learning Semantic Textual Similarity from Conversations
• Neural Graph Machines: Learning Neural Networks Using Graphs
• Reply With: Proactive Recommendation of Email Attachments
• A practical approach to dialogue response generation in closed domains
• Information Leakage in Embedding Models
• OCC: A Smart Reply System for Efficient In-App Communications
• Diversifying Reply Suggestions using a Matching-Conditional Variational Autoencoder
• A Repository of Conversational Datasets
• On the Social and Technical Challenges of Web Search Autosuggestion Moderation
• soc2seq: Social Embedding meets Conversation Model
• Netizen-Style Commenting on Fashion Photos: Dataset and Diversity Measures
• Soft Seeded SSL Graphs for Unsupervised Semantic Similarity-based Retrieval
• MoodSwipe: A Soft Keyboard that Suggests Messages Based on User-Specified Emotions
• Beyond "How may I help you?": Assisting Customer Service Agents with Proactive Responses
• MuTual: A Dataset for Multi-Turn Dialogue Reasoning
• Explicit and Implicit Semantic Ranking Framework
• MARVIN: An Open Machine Learning Corpus and Environment for Automated Machine Learning Primitive Annotation and Execution
• TEASPN: Framework and Protocol for Integrated Writing Assistance Environments
• Adversarial Semantic Collisions
• DeepNVM++: Cross-Layer Modeling and Optimization Framework of Non-Volatile Memories for Deep Learning
• Plan ahead: Self-Supervised Text Planning for Paragraph Completion Task
• A Bi-Encoder LSTM Model For Learning Unstructured Dialogs
• Improving Text Auto-Completion with Next Phrase Prediction
• Language Scaling for Universal Suggested Replies Model
• CryptoGRU: Low Latency Privacy-Preserving Text Analysis With GRU
• A Dataset and Baselines for Multilingual Reply Suggestion
• A Conditional Generative Matching Model for Multi-lingual Reply Suggestion