A Neural Network Approach to Context-Sensitive Generation of Conversational Responses
This work addresses the challenge of context-sensitive response generation for conversational AI, representing an incremental improvement over existing methods.
The paper tackles the problem of generating conversational responses by developing a neural network system that integrates contextual information from previous dialog utterances, trained on Twitter conversations, and shows consistent gains over baseline methods.
We present a novel response generation system that can be trained end to end on large quantities of unstructured Twitter conversations. A neural network architecture is used to address sparsity issues that arise when integrating contextual information into classic statistical models, allowing the system to take into account previous dialog utterances. Our dynamic-context generative models show consistent gains over both context-sensitive and non-context-sensitive Machine Translation and Information Retrieval baselines.