Ask Me Anything: Dynamic Memory Networks for Natural Language Processing

2015/06/24 by Ankit Kumar, Ozan İrsoy, Ozan Irsoy +18 · 2 voices · 631 citations
Computer Science · Psychology · #Artificial intelligence #Artificial neural network #Ask price #Cognition #Computer science #Episodic memory #Language model #Linguistics #Multimodal Machine Learning Applications #Natural (archaeology) #Natural Language Processing Techniques #Natural language #Natural language processing #Process (computing) #Programming language #Psychology #Question answering #Recurrent neural network #Sequence (biology) #Speech recognition #Topic Modeling #Treebank #Word (group theory) #cs.CL #cs.LG #cs.NE

paper · pdf · doi:10.48550/arxiv.1506.07285

published in arXiv (Cornell University) (Cornell University)

openalex publication_date 2015/06/24 · arxiv created 2016/03/05 · arxiv updated 2016/03/08 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

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.

Cited by

• Learning to Generate Reviews and Discovering Sentiment
• A Parallel-Hierarchical Model for Machine Comprehension on Sparse Data
• FVQA: Fact-Based Visual Question Answering
• HEAD-QA v2: Expanding a Healthcare Benchmark for Reasoning
• An Interpretable Reasoning Network for Multi-Relation Question Answering
• Gated Graph Sequence Neural Networks
• FVQA: Fact-based Visual Question Answering
• Stochastic Answer Networks for Machine Reading Comprehension
• Recent Advances in Recurrent Neural Networks
• Attention-Based Convolutional Neural Network for Machine Comprehension
• Ask Me Even More: Dynamic Memory Tensor Networks (Extended Model)
• Towards Neural Network-based Reasoning
• Motion-Appearance Co-Memory Networks for Video Question Answering
• Generative AI in Depth: A Survey of Recent Advances, Model Variants, and Real-World Applications
• Data Noising as Smoothing in Neural Network Language Models
• Aligned Image-Word Representations Improve Inductive Transfer Across Vision-Language Tasks
• Learning Semantically Coherent and Reusable Kernels in Convolution Neural Nets for Sentence Classification
• Separating Answers from Queries for Neural Reading Comprehension
• End-to-end Learning for Short Text Expansion
• Topic Modeling Using Distributed Word Embeddings
• Dual Attention Networks for Multimodal Reasoning and Matching
• Learning Phrase Embeddings from Paraphrases with GRUs
• A Read-Write Memory Network for Movie Story Understanding
• Skip-Thought Memory Networks
• Dual Attention Network for Product Compatibility and Function Satisfiability Analysis
• U-Net: Machine Reading Comprehension with Unanswerable Questions
• Neural Information Retrieval: A Literature Review
• Dialog state tracking, a machine reading approach using Memory Network
• Dynamic Adaptive Network Intelligence
• EKT: Exercise-Aware Knowledge Tracing for Student Performance Prediction
• Incorporating External Knowledge to Answer Open-Domain Visual Questions with Dynamic Memory Networks
• GILE: A Generalized Input-Label Embedding for Text Classification
• Pedestrian Tracking with Gated Recurrent Units and Attention Mechanisms
• Recognizing Implicit Discourse Relations via Repeated Reading: Neural Networks with Multi-Level Attention
• Neural Programmer: Inducing Latent Programs with Gradient Descent
• A Survey of Available Corpora for Building Data-Driven Dialogue Systems
• A Proposal for Intelligent Agents with Episodic Memory
• Label-semantics Aware Generative Approach for Domain-Agnostic Multilabel Classification
• Direct Prediction Set Minimization via Bilevel Conformal Classifier Training
• An Empirical Evaluation of Visual Question Answering for Novel Objects
• Recent Trends in Deep Learning Based Natural Language Processing
• Lie Access Neural Turing Machine
• The Goldilocks Principle: Reading Children's Books with Explicit Memory Representations
• No Need to Pay Attention: Simple Recurrent Neural Networks Work! (for Answering "Simple" Questions)
• Multilingual Hierarchical Attention Networks for Document Classification
• Representation Discrepancy Bridging Method for Remote Sensing Image-Text Retrieval
• Hierarchical Memory Networks
• Simultaneous Learning of Trees and Representations for Extreme Classification and Density Estimation
• Deep Residual Bidir-LSTM for Human Activity Recognition Using Wearable Sensors
• Learning better discourse representation for implicit discourse relation recognition via attention networks
• Automated identification of media bias in news articles: an interdisciplinary literature review
• CFO: Conditional Focused Neural Question Answering with Large-scale Knowledge Bases
• Visual Text Correction
• Image Denoising using Attention-Residual Convolutional Neural Networks

Discussions

• Ask Me Anything: Dynamic Memory Networks for Natural Language Processing [hn, 35 points, 5 comments]
• Ask Me Anything: Dynamic Memory Networks for Natural Language Processing [hn, 12 points, 2 comments]

Related

• Ask Me Even More: Dynamic Memory Tensor Networks (Extended Model)
• Dynamic Memory Networks for Visual and Textual Question Answering
• Question Dependent Recurrent Entity Network for Question Answering
• Learning to Compose Neural Networks for Question Answering
• Empirical Study on Deep Learning Models for Question Answering