Recurrent neural network language models (RNNLMs) have consistently outperformed n-gram language models when used in automatic speech recognition (ASR). This is because RNNLMs provide robust parameter estimation through the use of a continuous-space representation of words, and can generally model longer context dependencies than n-grams. The adaptation of RNNLMs to new domains remains an active research area and the two main approaches are: feature-based adaptation, where the input to the RNNLM is augmented with auxiliary features; and model-based adaptation, which includes model fine-tuning and introduction of adaptation layer(s) in the network. This paper explores the properties of both types of adaptation on multi-genre broadcast speech recognition. Two hybrid adaptation techniques are proposed, namely the finetuning of feature-based RNNLMs and the use of a feature-based adaptation layer. A method for the semi-supervised adaptation of RNNLMs, using topic model-based genre classification, is also presented and investigated. The gains obtained with RNNLM adaptation on a system trained on 700h. of speech are consistent using both RNNLMs trained on a small (10M words) and large set (660M words), with 10% perplexity and 2% word error rate improvements on a 28.3h. test set.

In speech pathology, new assistive technologies using ASR and machine learning approaches are being developed for detecting speech disorder events. Classically-trained ASR model tends to remove disfluencies from spoken utterances, due to its focus on producing clean and readable text output. However, diagnostic systems need to be able to track speech disfluencies, such as stuttering events, in order to determine the severity level of stuttering. To achieve this, ASR systems must be adapted to recognise full verbatim utterances, including pseudo-words and non-meaningful part-words. This work proposes a training regime to address this problem, and preserve a full verbatim output of stuttering speech. We use a lightly-supervised approach using task-oriented lattices to recognise the stuttering speech of children performing a standard reading task. This approach improved the WER by 27.8% relative to a baseline that uses word-lattices generated from the original prompt. The improved results preserved 63% of stuttering events (including sound, word, part-word and phrase repetition, and revision). This work also proposes a separate correction layer on top of the ASR that detects prolongation events (which are poorly recognised by the ASR). This increases the percentage of preserved stuttering events to 70%.

No abstract

BackgroundTools proposed to triage patient acuity in COVID-19 infection have only been validated in hospital populations. We estimated the accuracy of five risk-stratification tools recommended to predict severe illness and compared accuracy to existing clinical decision making in a prehospital setting.MethodsAn observational cohort study using linked ambulance service data for patients attended by Emergency Medical Service (EMS) crews in the Yorkshire and Humber region of England between 26 March 2020 and 25 June 2020 was conducted to assess performance of the Pandemic Respiratory Infection Emergency System Triage (PRIEST) tool, National Early Warning Score (NEWS2), WHO algorithm, CRB-65 and Pandemic Medical Early Warning Score (PMEWS) in patients with suspected COVID-19 infection. The primary outcome was death or need for organ support.ResultsOf the 7549 patients in our cohort, 17.6% (95% CI 16.8% to 18.5%) experienced the primary outcome. The NEWS2 (National Early Warning Score, version 2), PMEWS, PRIEST tool and WHO algorithm identified patients at risk of adverse outcomes with a high sensitivity (>0.95) and specificity ranging from 0.3 (NEWS2) to 0.41 (PRIEST tool). The high sensitivity of NEWS2 and PMEWS was achieved by using lower thresholds than previously recommended. On index assessment, 65% of patients were transported to hospital and EMS decision to transfer patients achieved a sensitivity of 0.84 (95% CI 0.83 to 0.85) and specificity of 0.39 (95% CI 0.39 to 0.40).ConclusionUse of NEWS2, PMEWS, PRIEST tool and WHO algorithm could improve sensitivity of EMS triage of patients with suspected COVID-19 infection. Use of the PRIEST tool would improve sensitivity of triage without increasing the number of patients conveyed to hospital.

No abstract

We describe the University of Sheffield system for participation in the 2015 Multi-Genre Broadcast (MGB) challenge task of transcribing multi-genre broadcast shows. Transcription was one of four tasks proposed in the MGB challenge, with the aim of advancing the state of the art of automatic speech recognition, speaker diarisation and automatic alignment of subtitles for broadcast media. Four topics are investigated in this work: Data selection techniques for training with unreliable data, automatic speech segmentation of broadcast media shows, acoustic modelling and adaptation in highly variable environments, and language modelling of multigenre shows. The final system operates in multiple passes, using an initial unadapted decoding stage to refine segmentation, followed by three adapted passes: a hybrid DNN pass with input features normalised by speaker-based cepstral normalisation, another hybrid stage with input features normalised by speaker feature-MLLR transformations, and finally a bottleneck-based tandem stage with noise and speaker factorisation. The combination of these three system outputs provides a final error rate of 27.5% on the official development set, consisting of 47 multi-genre shows.

No abstract

The evaluation of scientific submissions through peer review is both the most fundamental component of the publication process, as well as the most frequently criticised and questioned. Academic journals and conferences request reviews from multiple reviewers per submission, which an editor, or area chair aggregates into the final acceptance decision. Reviewers are often in disagreement due to varying levels of domain expertise, confidence, levels of motivation, as well as due to the heavy workload and the different interpretations by the reviewers of the score scale. Herein, we explore the possibility of a computational decision support tool for the editor, based on Natural Language Processing, that offers an additional aggregated recommendation. We provide a comparative study of state-of-the-art text modelling methods on the newly crafted, largest review dataset of its kind based on Interspeech 2019, and we are the first to explore uncertainty-aware methods (soft labels, quantile regression) to address the subjectivity inherent in this problem.