Knowledge Distillation in Document Retrieval
This work addresses the need for efficient document retrieval in fact extraction and verification, offering a practical solution for real-world applications, though it is incremental as it applies an existing technique to a specific domain.
The paper tackles the problem of creating fast, scalable document retrieval by using knowledge distillation to train claim-independent student models to mimic a complex claim-dependent teacher model, resulting in student models that are 12x faster and 20x smaller while improving ranking metrics.
Complex deep learning models now achieve state of the art performance for many document retrieval tasks. The best models process the query or claim jointly with the document. However for fast scalable search it is desirable to have document embeddings which are independent of the claim. In this paper we show that knowledge distillation can be used to encourage a model that generates claim independent document encodings to mimic the behavior of a more complex model which generates claim dependent encodings. We explore this approach in document retrieval for a fact extraction and verification task. We show that by using the soft labels from a complex cross attention teacher model, the performance of claim independent student LSTM or CNN models is improved across all the ranking metrics. The student models we use are 12x faster in runtime and 20x smaller in number of parameters than the teacher