End-to-End Open Vocabulary Keyword Search
This work addresses the challenge of keyword search in speech for applications requiring exact query localization, offering a solution that handles vocabulary openness and data imbalance, though it is incremental as it builds on existing neural approaches.
The authors tackled the problem of open vocabulary keyword search in speech by proposing a model that directly outputs frame-level probabilities for query occurrences, enabling exact location detection. Their model outperformed similar end-to-end models on balanced tasks and improved LVCSR-based systems through rescoring, achieving significant gains in keyword search performance.
Recently, neural approaches to spoken content retrieval have become popular. However, they tend to be restricted in their vocabulary or in their ability to deal with imbalanced test settings. These restrictions limit their applicability in keyword search, where the set of queries is not known beforehand, and where the system should return not just whether an utterance contains a query but the exact location of any such occurrences. In this work, we propose a model directly optimized for keyword search. The model takes a query and an utterance as input and returns a sequence of probabilities for each frame of the utterance of the query having occurred in that frame. Experiments show that the proposed model not only outperforms similar end-to-end models on a task where the ratio of positive and negative trials is artificially balanced, but it is also able to deal with the far more challenging task of keyword search with its inherent imbalance. Furthermore, using our system to rescore the outputs an LVCSR-based keyword search system leads to significant improvements on the latter.