Learning Semantic Textual Similarity from Conversations
arXiv:1804.07754
Abstract
We present a novel approach to learn representations for sentence-level semantic similarity using conversational data. Our method trains an unsupervised model to predict conversational input-response pairs. The resulting sentence embeddings perform well on the semantic textual similarity (STS) benchmark and SemEval 2017's Community Question Answering (CQA) question similarity subtask. Performance is further improved by introducing multitask training combining the conversational input-response prediction task and a natural language inference task. Extensive experiments show the proposed model achieves the best performance among all neural models on the STS benchmark and is competitive with the state-of-the-art feature engineered and mixed systems in both tasks.
arXiv comment10 pages, 8 Figures, 6 Tables
References in corpus (6)
- Unsupervised Learning of Sentence Embeddings using Compositional n-Gram Features
- A large annotated corpus for learning natural language inference
- Neural Natural Language Inference Models Enhanced with External Knowledge
- Smart Reply: Automated Response Suggestion for Email
- Conversational Contextual Cues: The Case of Personalization and History for Response Ranking
- Do latent tree learning models identify meaningful structure in sentences?
Cited by in corpus (12)
- ERNIE: Enhanced Representation through Knowledge Integration
- Learning semantic sentence representations from visually grounded language without lexical knowledge
- I love your chain mail! Making knights smile in a fantasy game world: Open-domain goal-oriented dialogue agents
- Effective Parallel Corpus Mining using Bilingual Sentence Embeddings
- Understanding Chat Messages for Sticker Recommendation in Messaging Apps
- Review-guided Helpful Answer Identification in E-commerce
- REGMAPR - Text Matching Made Easy
- An Efficient Approach for Geo-Multimedia Cross-Modal Retrieval
- Efficient Purely Convolutional Text Encoding
- A Case Study to Reveal if an Area of Interest has a Trend in Ongoing Tweets Using Word and Sentence Embeddings
- SentEmojiBot: Empathising Conversations Generation with Emojis
- Parameter-free Sentence Embedding via Orthogonal Basis