Efficient Retrieval Augmented Generation from Unstructured Knowledge for Task-Oriented Dialog
This work provides methods for more efficient knowledge retrieval and response generation in task-oriented dialog systems, which is important for developers building such systems.
This paper addresses the challenge of generating responses in task-oriented dialogs that require knowledge from unstructured documents. The authors propose a hierarchical classification approach for knowledge selection, achieving their best results. Alternatively, a siamese sequence embedding model (Dense Knowledge Retrieval) reduces computation time by over 100x, albeit with a 5-6% degradation in R@1.
This paper summarizes our work on the first track of the ninth Dialog System Technology Challenge (DSTC 9), "Beyond Domain APIs: Task-oriented Conversational Modeling with Unstructured Knowledge Access". The goal of the task is to generate responses to user turns in a task-oriented dialog that require knowledge from unstructured documents. The task is divided into three subtasks: detection, selection and generation. In order to be compute efficient, we formulate the selection problem in terms of hierarchical classification steps. We achieve our best results with this model. Alternatively, we employ siamese sequence embedding models, referred to as Dense Knowledge Retrieval, to retrieve relevant documents. This method further reduces the computation time by a factor of more than 100x at the cost of degradation in R@1 of 5-6% compared to the first model. Then for either approach, we use Retrieval Augmented Generation to generate responses based on multiple selected snippets and we show how the method can be used to fine-tune trained embeddings.