Semantically Conditioned Dialog Response Generation via Hierarchical Disentangled Self-Attention
This work addresses the problem of scaling dialog systems to handle multiple domains efficiently for developers and researchers, representing an incremental advancement with a novel method for a known bottleneck.
The paper tackles the scalability issue in semantically controlled neural response generation for multi-domain large-scale scenarios by introducing a hierarchical disentangled self-attention network that models dialog acts as a graph, resulting in significant improvements on the Multi-Domain-WOZ dataset across automatic and human evaluation metrics.
Semantically controlled neural response generation on limited-domain has achieved great performance. However, moving towards multi-domain large-scale scenarios are shown to be difficult because the possible combinations of semantic inputs grow exponentially with the number of domains. To alleviate such scalability issue, we exploit the structure of dialog acts to build a multi-layer hierarchical graph, where each act is represented as a root-to-leaf route on the graph. Then, we incorporate such graph structure prior as an inductive bias to build a hierarchical disentangled self-attention network, where we disentangle attention heads to model designated nodes on the dialog act graph. By activating different (disentangled) heads at each layer, combinatorially many dialog act semantics can be modeled to control the neural response generation. On the large-scale Multi-Domain-WOZ dataset, our model can yield a significant improvement over the baselines on various automatic and human evaluation metrics.