Training language models for Code-mixed (CM) language is known to be a difficult problem
because of lack of data compounded by the increased confusability due to the presence of …

In this paper, we report the development of the South East Asia Mandarin-English (SEAME)
corpus, including 63 hours of transcribed spontaneous Mandarin-English code-switching …

LSTM based language models are an important part of modern LVCSR systems as they
significantly improve performance over traditional backoff language models. Incorporating …

Curriculum learning strategies are known to improve the accuracy, robustness and
convergence rate for various language learning tasks using deep architectures (Bengio et …

Air Navigation Service Providers (ANSPs) replace paper flight strips through different digital
solutions. The instructed commands from an air traffic controller (ATCos) are then available …

Although neural language models have emerged, n-gram language models are still used for
many speech recognition tasks. This paper proposes four methods to improve n-gram …

Abstract Traditionally, short-range Language Models (LMs) like the conventional n-gram
models have been used for language model adaptation. Recent work has improved …

Although n-gram language models (LMs) have been outperformed by the state-of-the-art
neural LMs, they are still widely used in speech recognition due to its high efficiency in …

Abstract Automatic Speech Recognition (ASR) has recently proved to be a useful tool to
reduce the workload of air traffic controllers leading to significant gains in operational …

Large pre-trained language models (PLMs) have shown remarkable performance across
various natural language understanding (NLU) tasks, particularly in low-resource settings …