Deep Similarity Neural Network QA System
Jump to navigation
Jump to search
A Deep Similarity Neural Network QA System is a QA System that is based on a Deep Similarity Neural Network model.
- Example(s):
- Counter-Example(s):
- See: Semantic Similarity Neural Network, Artificial Neural Network, Deep Learning Neural Network, Natural Language Processing Task, Costumer Care Chat System.
References
2017
- (Minaee & Liu, 2017) ⇒ Shervin Minaee, and Zhu Liu (2017). "Automatic question-answering using a deep similarity neural network". arXiv preprint arXiv:1708.01713.
- QUOTE: In this work we propose a model for question-answering using a deep similarity network. First, a neural network based model is used to embed the question and answer into a low-dimensional representation. Then, these features are fed into two parallel neural networks, and combined after a few layers of hierarchical representation to make the final decision.
The main target application of this model is for AT&T customer care chat system, where we use the proposed model to find the best answer from the database of old chat data, and find the confidence score of the relatedness of the selected answer for the current question. If the confidence score is larger than some threshold, it would be retrieved to the customer, otherwise the question will be referred to the human agent. In order to keep the users’ identities and personal information protected, all the chat data used in this paper have been anonymized.
- QUOTE: In this work we propose a model for question-answering using a deep similarity network. First, a neural network based model is used to embed the question and answer into a low-dimensional representation. Then, these features are fed into two parallel neural networks, and combined after a few layers of hierarchical representation to make the final decision.