Recurrent Convolutional Neural Networks for Discourse Compositionality
arXiv:1306.3584
Abstract
The compositionality of meaning extends beyond the single sentence. Just as words combine to form the meaning of sentences, so do sentences combine to form the meaning of paragraphs, dialogues and general discourse. We introduce both a sentence model and a discourse model corresponding to the two levels of compositionality. The sentence model adopts convolution as the central operation for composing semantic vectors and is based on a novel hierarchical convolutional neural network. The discourse model extends the sentence model and is based on a recurrent neural network that is conditioned in a novel way both on the current sentence and on the current speaker. The discourse model is able to capture both the sequentiality of sentences and the interaction between different speakers. Without feature engineering or pretraining and with simple greedy decoding, the discourse model coupled to the sentence model obtains state of the art performance on a dialogue act classification experiment.
References in corpus (1)
Cited by in corpus (30)
- A Convolutional Neural Network for Modelling Sentences
- Neural Summarization by Extracting Sentences and Words
- Multilingual Models for Compositional Distributed Semantics
- Multilingual Distributed Representations without Word Alignment
- Multimodal Convolutional Neural Networks for Matching Image and Sentence
- A Comparison of Machine Learning Algorithms for the Surveillance of Autism Spectrum Disorder
- Named Entity Recognition with stack residual LSTM and trainable bias decoding
- Conversational Analysis using Utterance-level Attention-based Bidirectional Recurrent Neural Networks
- Modeling Long-Range Context for Concurrent Dialogue Acts Recognition
- A Deep Architecture for Semantic Parsing
- Distributed Representations for Compositional Semantics
- English-Bhojpuri SMT System: Insights from the Karaka Model
- WikiReading: A Novel Large-scale Language Understanding Task over Wikipedia
- Abstractive Dialogue Summarization with Sentence-Gated Modeling Optimized by Dialogue Acts
- Discourse-Wizard: Discovering Deep Discourse Structure in your Conversation with RNNs
- Hierarchical Multi-Label Dialog Act Recognition on Spanish Data
- TextNAS: A Neural Architecture Search Space tailored for Text Representation
- Joint Modeling of Content and Discourse Relations in Dialogues
- How did the discussion go: Discourse act classification in social media conversations
- Convolutional Neural Networks with Recurrent Neural Filters
- Speaker-change Aware CRF for Dialogue Act Classification
- Basic tasks of sentiment analysis
- Dialogue Act Classification in Group Chats with DAG-LSTMs
- DCR-Net: A Deep Co-Interactive Relation Network for Joint Dialog Act Recognition and Sentiment Classification
- Improved Dynamic Memory Network for Dialogue Act Classification with Adversarial Training
- Resolving Lexical Ambiguity in Tensor Regression Models of Meaning
- Speeding Up Neural Machine Translation Decoding by Cube Pruning
- A Deep Learning-Based Approach for Measuring the Domain Similarity of Persian Texts
- Unsupervised Dialogue Act Induction using Gaussian Mixtures
- Compositional Sentence Representation from Character within Large Context Text