Ask Me Anything: Dynamic Memory Networks for Natural Language Processing

Abstract

Most tasks in natural language processing can be cast into question answering (QA) problems over language input. We introduce the dynamic memory network (DMN), a neural network architecture which processes input sequences and questions, forms episodic memories, and generates relevant answers. Questions trigger an iterative attention process which allows the model to condition its attention on the inputs and the result of previous iterations. These results are then reasoned over in a hierarchical recurrent sequence model to generate answers. The DMN can be trained end-to-end and obtains state-of-the-art results on several types of tasks and datasets: question answering (Facebook's bAbI dataset), text classification for sentiment analysis (Stanford Sentiment Treebank) and sequence modeling for part-of-speech tagging (WSJ-PTB). The training for these different tasks relies exclusively on trained word vector representations and input-question-answer triplets.

References in corpus (9)

• Sequence to Sequence Learning with Neural Networks
• Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
• Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
• Distributed Representations of Sentences and Documents
• Teaching Machines to Read and Comprehend
• A Multi-World Approach to Question Answering about Real-World Scenes based on Uncertain Input
• Neural GPUs Learn Algorithms
• Learning a Recurrent Visual Representation for Image Caption Generation
• Inferring Algorithmic Patterns with Stack-Augmented Recurrent Nets

Cited by in corpus (80)

• Recent Trends in Deep Learning Based Natural Language Processing
• Recent Advances in Recurrent Neural Networks
• The Goldilocks Principle: Reading Children's Books with Explicit Memory Representations
• Gated Graph Sequence Neural Networks
• Learning to Generate Reviews and Discovering Sentiment
• ReasoNet: Learning to Stop Reading in Machine Comprehension
• An overview and comparative analysis of Recurrent Neural Networks for Short Term Load Forecasting
• Visual Question Answering: Datasets, Algorithms, and Future Challenges
• A Survey of Available Corpora for Building Data-Driven Dialogue Systems
• Data Noising as Smoothing in Neural Network Language Models
• The NLP Cookbook: Modern Recipes for Transformer based Deep Learning Architectures
• Lifelong Sequential Modeling with Personalized Memorization for User Response Prediction
• 3-D Scene Graph: A Sparse and Semantic Representation of Physical Environments for Intelligent Agents
• An Interpretable Reasoning Network for Multi-Relation Question Answering
• Aspect Level Sentiment Classification with Deep Memory Network
• Neural Programmer: Inducing Latent Programs with Gradient Descent
• Towards Neural Network-based Reasoning
• Asymmetric Residual Neural Network for Accurate Human Activity Recognition
• Hierarchical Memory Networks
• Multilingual Hierarchical Attention Networks for Document Classification
• Dual Attention Networks for Multimodal Reasoning and Matching
• Neural Information Retrieval: A Literature Review
• Incorporating External Knowledge to Answer Open-Domain Visual Questions with Dynamic Memory Networks
• Generative Temporal Models with Memory
• U-Net: Machine Reading Comprehension with Unanswerable Questions
• FVQA: Fact-based Visual Question Answering
• Compositional Memory for Visual Question Answering
• Stochastic Answer Networks for Machine Reading Comprehension
• A Taxonomy for Neural Memory Networks
• Visual Relationship Detection with Visual-Linguistic Knowledge from Multimodal Representations
• CFO: Conditional Focused Neural Question Answering with Large-scale Knowledge Bases
• Results of the seventh edition of the BioASQ Challenge
• Sentence Pair Scoring: Towards Unified Framework for Text Comprehension
• A Parallel-Hierarchical Model for Machine Comprehension on Sparse Data
• Motion-Appearance Co-Memory Networks for Video Question Answering
• Deep Residual Bidir-LSTM for Human Activity Recognition Using Wearable Sensors
• A Comprehensive Review of State-of-The-Art Methods for Java Code Generation from Natural Language Text
• Combining Convolution and Recursive Neural Networks for Sentiment Analysis
• Neural Discourse Relation Recognition with Semantic Memory
• Image Denoising using Attention-Residual Convolutional Neural Networks
• Finding Answers from the Word of God: Domain Adaptation for Neural Networks in Biblical Question Answering
• Learning in Text Streams: Discovery and Disambiguation of Entity and Relation Instances
• Separating Answers from Queries for Neural Reading Comprehension
• A Read-Write Memory Network for Movie Story Understanding
• Learning with Memory Embeddings
• Simultaneous Learning of Trees and Representations for Extreme Classification and Density Estimation
• Dialog state tracking, a machine reading approach using Memory Network
• Attention-Based Convolutional Neural Network for Machine Comprehension
• Representation Stability as a Regularizer for Improved Text Analytics Transfer Learning
• No Need to Pay Attention: Simple Recurrent Neural Networks Work! (for Answering "Simple" Questions)
• Visual Text Correction
• Lie Access Neural Turing Machine
• An Empirical Evaluation of Visual Question Answering for Novel Objects
• Sparse Attentive Memory Network for Click-through Rate Prediction with Long Sequences
• Recognizing Implicit Discourse Relations via Repeated Reading: Neural Networks with Multi-Level Attention
• A Neural Architecture Mimicking Humans End-to-End for Natural Language Inference
• Combination of Domain Knowledge and Deep Learning for Sentiment Analysis of Short and Informal Messages on Social Media
• Learning Phrase Embeddings from Paraphrases with GRUs
• Evolutionary Training and Abstraction Yields Algorithmic Generalization of Neural Computers
• Learning Semantically Coherent and Reusable Kernels in Convolution Neural Nets for Sentence Classification
• The Tensor Brain: Semantic Decoding for Perception and Memory
• Sentiment Analysis with Contextual Embeddings and Self-Attention
• Topic Modeling Using Distributed Word Embeddings
• A Language-independent and Compositional Model for Personality Trait Recognition from Short Texts
• Reading Comprehension using Entity-based Memory Network
• Ask Me Even More: Dynamic Memory Tensor Networks (Extended Model)
• GILE: A Generalized Input-Label Embedding for Text Classification
• Extending Compositional Attention Networks for Social Reasoning in Videos
• Separate and Attend in Personal Email Search
• Memory-guided Unsupervised Image-to-image Translation
• Aligned Image-Word Representations Improve Inductive Transfer Across Vision-Language Tasks
• Hierarchical Attention Model for Improved Machine Comprehension of Spoken Content
• Skip-Thought Memory Networks
• Dynamic Adaptive Network Intelligence
• A Proposal for Intelligent Agents with Episodic Memory
• A Hierarchical Approach to Neural Context-Aware Modeling
• End-to-end Learning for Short Text Expansion
• Here's My Point: Joint Pointer Architecture for Argument Mining
• Pedestrian Tracking with Gated Recurrent Units and Attention Mechanisms
• Dual Attention Network for Product Compatibility and Function Satisfiability Analysis