Probing Statistical Representations For End-To-End ASR
This work provides analysis for improving ASR models and downstream tasks, but it is incremental as it builds on existing methods without major breakthroughs.
The paper investigates cross-domain language model dependencies within transformer architectures for end-to-end ASR using SVCCA analysis, finding that specific neural representations exhibit correlated behavior impacting recognition performance.
End-to-End automatic speech recognition (ASR) models aim to learn a generalised speech representation to perform recognition. In this domain there is little research to analyse internal representation dependencies and their relationship to modelling approaches. This paper investigates cross-domain language model dependencies within transformer architectures using SVCCA and uses these insights to exploit modelling approaches. It was found that specific neural representations within the transformer layers exhibit correlated behaviour which impacts recognition performance. Altogether, this work provides analysis of the modelling approaches affecting contextual dependencies and ASR performance, and can be used to create or adapt better performing End-to-End ASR models and also for downstream tasks.