Improve Speech Recognition Research Articles

Some Challenging Questions About Outcomes in Children With Cochlear Implants

Purpose: Cochlear implants (CIs) have improved the quality of life for many children with severe-to-profound sensorineural hearing loss. Despite the reported CI benefits of improved speech recognition, speech intelligibility, and spoken language processing, large individual differences in speech and language outcomes are still consistently reported in the literature. The enormous variability in CI outcomes has made it challenging to predict which children may be at high risk for limited benefits and how potential risk factors can be improved with interventions. In this article, we consider four challenging, unresolved issues that exist in the field of pediatric cochlear implantation (age at implantation, comorbid diagnoses, preimplant functioning in infants, and communication modality) to illustrate the value of adopting a broader integrated multidisciplinary approach to explaining variability and individual differences in outcomes. Conclusions: Four challenging and unresolved issues concerning our ability to predict individual differences before implantation were discussed. In order to address some of the current barriers to progress in dealing with these unresolved issues, a biopsychosocial system model is proposed as a possible unifying approach to conceptualizing these issues. Additionally, several new research directions to better understand and explain variability in pediatric CI outcomes are suggested.

Implementation of a Whisper Architecture-Based Turkish Automatic Speech Recognition (ASR) System and Evaluation of the Effect of Fine-Tuning with a Low-Rank Adaptation (LoRA) Adapter on Its Performance

This paper focuses on the implementation of the Whisper architecture to create an automatic speech recognition (ASR) system optimized for the Turkish language, which is considered a low-resource language in terms of speech recognition technologies. Whisper is a transformer-based model known for its high performance across numerous languages. However, its performance in Turkish, a language with unique linguistic features and limited labeled data, has yet to be fully explored. To address this, we conducted a series of experiments using five different Turkish speech datasets to assess the model’s baseline performance. Initial evaluations revealed a range of word error rates (WERs) between 4.3% and 14.2%, reflecting the challenges posed by Turkish. To improve these results, we applied the low-rank adaptation (LoRA) technique, which is designed to fine-tune large-scale models efficiently by introducing a reduced set of trainable parameters. After fine-tuning, significant performance improvements were observed, with WER reductions of up to 52.38%. This study demonstrates that fine-tuned Whisper models can be successfully adapted for Turkish, resulting in a robust and accurate end-to-end ASR system. This research highlights the applicability of Whisper in low-resource languages and provides insights into the challenges of and strategies for improving speech recognition performance in Turkish.

Directionality in BiCROS hearing aids: an investigation of objective and subjective outcomes

Objectives To assess the effect of forward and omnidirectional microphone configurations in BiCROS versus monaural hearing aids on objective and subjective outcomes in different noise conditions. Design After fitting and a 4-week acclimatisation period, speech recognition and sound quality were measured using forward directional, omnidirectional, and unaided settings. Two noise configurations were used, surrounding noise and noise presented from the aided (better) ear. Subjective outcomes were assessed using the SSQ-b and BBSS questionnaires and participant interviews. Study sample Eighteen adult participants (mean: 74.6 y; range: 61–94 y; ten males, eight females) with mild to moderately severe SNHL in their better ear (PTA0.5–4khz > 20 dB HL) and limited usable hearing in their poorer ear (average PTA0.5–4khz > 100 dB HL). Results Significant improvement in speech recognition and sound quality for BiCROS and monaural directional settings over omnidirectional and unaided, in both noise configurations. There were no significant differences observed between monoaural and BiCROS directional settings. Conclusion Speech in noise recognition and sound quality scores demonstrated a significant directional benefit for both BiCROS and monaural directional fitting settings over omnidirectional and unaided conditions. Unique BiCROS-specific experiences were identified in a patient-oriented approach. These can inform the development of BiCROS-tailored tools.

Hearing What You Can't See: Influence of Face Masks on Speech Perception and Eye Movement by Adults With Hearing Loss.

The aim of the study was to analyze how face masks influence speech perception and time spent looking at the speaker's mouth and eyes by adults with and without hearing loss. Twenty participants with hearing loss and 20 without were asked to repeat Spanish words presented in various conditions, including different types of face masks (no mask, transparent window mask, and opaque mask FFP2) and presentation modes (audiovisual, video only, and audio only). Recognition accuracy and the percentage of time looking at the speaker's eyes and mouth (dwell time) were measured. In the audiovisual condition, participants with hearing loss had significantly better word recognition scores when the speaker wore no mask compared to when they wore an opaque face mask. However, there were no differences between the transparent mask and no mask conditions. For those with typical hearing, the type of face mask did not affect speech recognition. Audiovisual presentation consistently improved speech recognition for participants with hearing loss across all face mask conditions, but for those with typical hearing, it only improved compared to video-only mode. These participants demonstrated a ceiling effect in audiovisual and audio-only modes. Regarding eye movement patterns, participants spent less time looking at the speaker's mouth and more time at the eyes when the speaker wore an opaque mask compared to no mask or a transparent mask. The use of transparent face masks (ClearMask-type model) is recommended in contexts where face masks are still used (hospitals) to prevent the hindering effect of opaque masks (FFP2-type model) in speech perception among people with hearing loss, provided that any fogging of the window of the transparent mask is controlled by wiping it off as needed and the light is in front of the speaker to minimize shadows.

Interference of mid-level sound statistics underlie human speech recognition sensitivity in natural noise.

Recognizing speech in noise, such as in a busy restaurant, is an essential cognitive skill where the task difficulty varies across environments and noise levels. Although there is growing evidence that the auditory system relies on statistical representations for perceiving 1-5 and coding4,6-9 natural sounds, it's less clear how statistical cues and neural representations contribute to segregating speech in natural auditory scenes. We demonstrate that human listeners rely on mid-level statistics to segregate and recognize speech in environmental noise. Using natural backgrounds and variants with perturbed spectro-temporal statistics, we show that speech recognition accuracy at a fixed noise level varies extensively across natural backgrounds (0% to 100%). Furthermore, for each background the unique interference created by summary statistics can mask or unmask speech, thus hindering or improving speech recognition. To identify the neural coding strategy and statistical cues that influence accuracy, we developed generalized perceptual regression, a framework that links summary statistics from a neural model to word recognition accuracy. Whereas a peripheral cochlear model accounts for only 60% of perceptual variance, summary statistics from a mid-level auditory midbrain model accurately predicts single trial sensory judgments, accounting for more than 90% of the perceptual variance. Furthermore, perceptual weights from the regression framework identify which statistics and tuned neural filters are influential and how they impact recognition. Thus, perception of speech in natural backgrounds relies on a mid-level auditory representation involving interference of multiple summary statistics that impact recognition beneficially or detrimentally across natural background sounds.

Post-Stroke Dysarthria Voice Recognition based on Fusion Feature MSA and 1D

Post-stroke Dysarthria (PSD) is one of the common sequelae of stroke. PSD can harm patients’ quality of life and, in severe cases, be life-threatening. Most of the existing methods use frequency domain features to recognize the pathological voice, which makes it hard to completely represent the characteristics of pathological voice. Although some results have been achieved, there is still a long way to go for practical applications. Therefore, an improved deep learning-based model is proposed to classify between the pathological voice and the normal voice, using a novel fusion feature (MSA) and an improved 1D ResNet network hybrid bi-directional LSTM with dilated convolution (named 1D DRN-biLSTM). The experimental results show that our fusion features bring greater improvement in pathological speech recognition than the method that only analyzes the MFCC features, and can better synthesize the hidden features that characterize pathological speech. In terms of model structure, the introduction of dilated convolution and LSTM can further improve the performance of the 1D Resnet network, compared to ordinary networks such as CNN and LSTM. The accuracy of this method reaches 82.41% and 100% at the syllable level and speaker level, respectively. Our scheme outperforms other existing methods in terms of feature learning capability and recognition rate, and will help to play an important role in the assessment and diagnosis of PSD in China.

Phoneme Recognition in Korean Singing Voices Using Self-Supervised English Speech Representations

In general, it is difficult to obtain a huge, labeled dataset for deep learning-based phoneme recognition in singing voices. Studying singing voices also offers inherent challenges, compared to speech, because of the distinct variations in pitch, duration, and intensity. This paper proposes a detouring method to overcome this insufficient dataset, and applies it to the recognition of Korean phonemes in singing voices. The method started with pre-training the HuBERT, a self-supervised speech representation model, on a large-scale English corpus. The model was then adapted to the Korean speech domain with a relatively small-scale Korean corpus, in which the Korean phonemes were interpreted as similar English ones. Finally, the speech-adapted model was again trained with a tiny-scale Korean singing voice corpus for speech–singing adaptation. In the final adaptation, melodic supervision was chosen, which utilizes pitch information to improve the performance. For evaluation, the performance on multi-level error rates based on Word Error Rate (WER) was taken. Using the HuBERT-based transfer learning for adaptation improved the phoneme-level error rate of Korean speech by as much as 31.19%. Again, on singing voices by melodic supervision, it improved the rate by 0.55%. The significant improvement in speech recognition underscores the considerable potential of a model equipped with general human voice representations captured from the English corpus that can improve phoneme recognition on less target speech data. Moreover, the musical variation in singing voices is beneficial for phoneme recognition in singing voices. The proposed method could be applied to the phoneme recognition of other languages that have less speech and singing voice corpora.

Evaluating cochlear implant outcomes in DFNA9 subjects: a comprehensive study on cerebral white matter lesions and vestibular abnormalities.

This study assessed whether the Fazekas score could account for the variability in cochlear implantation (CI) outcomes among individuals with DFNA9 and evaluated signal loss in the semicircular canals (SCCs) on magnetic resonance imaging (MRI) among individuals with DFNA9. This retrospective cross-sectional study included CI recipients with DFNA9. Pre-implantation MRI-scans were reviewed to determine the Fazekas score, localizing and grading cerebral white matter lesions (WML), and identify abnormalities in the SCCs. CI performance was assessed by evaluating phoneme scores one year post-implantation. The function of the SCCs was evaluated using rotatory chair testing with electronystagmography (ENG) and the video Head Impulse Test (vHIT). Forty-five subjects (49 ears) were enrolled. The phoneme scores significantly improved from 35% (IQR 11-50) pre-implantation to 84% (IQR 76-90) one year post-implantation. No correlation was observed between the Fazekas score and the one-year post-implantation phoneme score (rsp=0.003, p = 0.986). Signal loss in at least one SCCs was detected in 97.7% of subjects and 77.8% of ears. There was no correlation between vestibular test results and fluid signal loss in the SCCs on MRI. Most individuals with DFNA9 show improved speech recognition with CI. The observed variability in CI outcomes was not linked to the Fazekas score. Additionally, our study confirms a high prevalence of focal sclerosis in DFNA9. Recognizing the limitations of this study, further research is needed to explore the predictive role of the Fazekas score on CI outcomes and its relationship with vestibular function.

Benefits of Cochlear Implantation for Older Adults With Asymmetric Hearing Loss.

FDA-approved indications for cochlear implantation include patients with severe-to-profound unilateral hearing loss (UHL) or asymmetric hearing loss (AHL); however, these indications are not covered for Medicare beneficiaries. We assessed the outcomes of cochlear implant (CI) use for older adults with UHL or AHL. Eighteen older adults (≥65 years of age at surgery) with UHL/AHL participated in a prospective, longitudinal investigation evaluating outcomes of CI use. Speech recognition for the affected ear was evaluated with consonant-nucleus-consonant (CNC) words. Spatial hearing was assessed with measures of sound source localization and sentence recognition in noise. The target sentence was presented from the front and the masker was either co-located with the target (SoNo), presented toward the affected ear (SoNci) or contralateral ear (SoNcontra). Perceived benefit was assessed with the Speech, Spatial, and Qualities of Hearing scale (SSQ) and the Tinnitus Handicap Inventory (THI). Participants experienced significant improvements with CI use for CNC words (mean [SD]; preop: 8% [10%], 1 yr: 51% [22%], 5 yr: 50% [19%]), masked sentence recognition (SoNcontra preop: 5% [6%], 1 yr: 22% [15%], 5 yr: 41% [14%]), and localization (preop: 76° [18°], 1 yr: 40° [11°], 5 yr: 41° [14°]),and reported significant improvements in hearing abilities (SSQ Spatial Hearing preop: 3 [1], 1 yr: 6 [2], 5 yr: 6 [2]) and tinnitus severity (THI preop: 16 [18], 1 yr: 4 [14], 5 yr: 6 [12]). Older adults with UHL/AHL experience significant improvements in speech recognition, spatial hearing, and subjective perceptions (e.g., hearing abilities and tinnitus severity) with a CI as compared to pre-operative abilities. IV Laryngoscope, 2024.

A Study on the Impact of Voice-to-Text Technology on Academic Achievement of the Hearing-Impaired

Hearing loss is a significant barrier to academic achievement, with hearing-impaired (HI) individuals often facing challenges in speech recognition, language development, and social interactions. Lip-reading, a crucial skill for HI individuals, is essential for effective communication and learning. However, the COVID-19 pandemic has exacerbated the challenges faced by HI individuals, with the face masks hindering lip-reading. This literature review explores the relationship between hearing loss and academic achievement, highlighting the importance of lip-reading and the potential of artificial intelligence (AI) techniques in mitigating these challenges. The introduction of Voice-to-Text (VtT) technology, which provides real-time text captions, can significantly improve speech recognition and academic performance for HI students. AI models, such as Hidden Markov models and Transformer models, can enhance the accuracy and robustness of VtT technology in diverse educational settings. Furthermore, VtT technology can facilitate better teacher-student interactions, provide transcripts of lectures and classroom discussions, and bridge the gap in standardized testing performance between HI and hearing students. While challenges and limitations exist, the successful implementation of VtT technology can promote inclusive education and enhance academic achievement. Future research directions include popularizing VtT technology, addressing technological barriers, and customizing VtT systems to cater to individual needs.

Speech Recognition and Subjective Hearing Abilities for Electric-Acoustic Stimulation Users With Unilateral Hearing Loss.

Some cochlear implant (CI) recipients with unilateral hearing loss (UHL) have functional acoustic hearing in the implanted ear, warranting the fitting of an ipsilateral electric-acoustic stimulation (EAS) device. The present study assessed speech recognition and subjective hearing abilities over time for EAS users with UHL. Prospective, repeated-measures. Tertiary referral center. Adult CI recipients with normal-to-moderate low-frequency acoustic thresholds in the implanted ear and a contralateral pure-tone average (0.5, 1, and 2 kHz) ≤25 dB HL. Participants were evaluated preoperatively and at 1, 3, and 6 months post-activation. Speech recognition for the affected ear was evaluated with CNC words in quiet. Masked speech recognition in the bilateral condition was evaluated with AzBio sentences in a 10-talker masker (0 dB SNR) for three spatial configurations: target from the front and masker either colocated with the target or presented 90 degrees toward the affected or contralateral ear. Responses to the Speech, Spatial, and Qualities of Hearing Scale subscales were used to assess subjective hearing abilities. Participants experienced significant improvements in CNC scores ( F(3,13) = 14.90, p < 0.001), and masked speech recognition in the colocated ( F(3,11) = 3.79, p = 0.043) and masker toward the contralateral ear ( F(3,11) = 4.75, p = 0.023) configurations. They also reported significantly better abilities on the Speech Hearing ( F(3,13) = 5.19, p = 0.014) and Spatial Hearing ( F(3,13) = 10.22, p = 0.001) subscales. Adults with UHL and functional acoustic hearing in the implanted ear experience significant improvements in speech recognition and subjective hearing abilities within the initial months of EAS use as compared with preoperative performance and perceptions.

Low Resource Malay Dialect Automatic Speech Recognition Modeling Using Transfer Learning from a Standard Malay Model

Approaches to automatic speech recognition have transited from Hidden Markov Model (HMM)-based ASR to deep neural networks. The advantages of deep neural network approaches are that they can be developed quickly and perform better given large language resources. Nevertheless, dialect speech recognition is still challenging due to the limited resources. Transfer learning approaches have been proposed to improve speech recognition for low resources. In the first approach, the model is pre-trained on a large and diverse labeled dataset to learn the acoustic and language patterns from the speech signal. Then, the model parameters are updated with a new dataset, and the pre-trained model is fine-tuned on a low-resource language dataset. The fine-tuning process is usually completed by freezing the pre-trained layers and training the remaining layers of the model on the low-resource language corpus. Another approach is to use a pre-trained model to capture the compact and meaningful features as input to the encoder. Pre-training in this approach usually involves using unsupervised learning methods to train models on a corpus of large amounts of unmarked data. It enables the model to learn the general patterns and relationships between the input speech signals. This paper proposes a training recipe using transfer learning and Standard Malay models to improve automatic speech recognition for Kelantan and Sarawak Malay dialects.

Implementation of Sound Direction Detection and Mixed Source Separation in Embedded Systems.

In recent years, embedded system technologies and products for sensor networks and wearable devices used for monitoring people's activities and health have become the focus of the global IT industry. In order to enhance the speech recognition capabilities of wearable devices, this article discusses the implementation of audio positioning and enhancement in embedded systems using embedded algorithms for direction detection and mixed source separation. The two algorithms are implemented using different embedded systems: direction detection developed using TI TMS320C6713 DSK and mixed source separation developed using Raspberry Pi 2. For mixed source separation, in the first experiment, the average signal-to-interference ratio (SIR) at 1 m and 2 m distances was 16.72 and 15.76, respectively. In the second experiment, when evaluated using speech recognition, the algorithm improved speech recognition accuracy to 95%.

Middle Ear Fibrosis Contributes to Hearing Loss in Patients With Myhre Syndrome.

Myhre syndrome (MS) is a rare genetic condition that presents with multiple genetic anomalies including cleft lip and palate and Eustachian tube dysfunction. These patients are at a high risk for airway scarring from intubation and mucosal inflammation. Hearing loss (conductive or mixed, of varying severity) is a common comorbidity in these patients, the exact etiology of which is still unclear. We present the cases of 2 unrelated children with MS who suffered progressive mixed hearing loss from fibrosis and obliteration of the middle ear spaces. Both patients had multiple sets of ear tubes that demonstrated early extrusion. The older patient underwent bone conduction implantation at age 11 which resulted in dramatic improvement of speech recognition and interactive skills. The other younger patient demonstrates a similar trajectory but has not yet undergone implantation. Otolaryngologists should take a cautious approach to surgery of the eardrum and middle ear to avoid unnecessary induction of fibrosis in this susceptible patient population. These cases highlight a newly described etiology for hearing loss and suggest a benefit to bone conduction implantation.

Personalized recommendation based on improved speech recognition algorithm in music e-learning course simulation

In recent years, the rapid development of internet technology has promoted the integration of music performance training systems with new teaching environments. However, personalized recommendations related to users are still a drawback of traditional performance training. Therefore, this article optimizes the system based on the shortcomings of traditional performance training systems and adds algorithms related to speech recognition and personalized recommendations to the music performance training system. Traditional music performance systems still have a lot of problems, resulting in a narrow application prospect. Therefore, this article first improves the speech recognition algorithm, and then uses the advantages of personalized recommendation algorithms in filtering applications to reconstruct the music performance training system. This article has conducted a large number of testing experiments on the system, and the experimental results strongly prove that the optimized speech recognition algorithm is superior to the algorithms used in traditional performance training systems. The model of this research not only incorporates personalized recommendations into the system, but also undergoes multiple optimizations in the experiment, ultimately obtaining a satisfactory product. The system can provide a suitable training platform for music performance, Effectively changing the playing environment and making music performance training more scientific and digital.

Failure in HiRes Ultra Series Recall Devices Does Not Necessarily Lead to Decrement in Performance.

To understand the impact on speech perception for patients experiencing Advanced Bionics V1 series Ultra and Ultra 3D cochlear implant failure. Retrospective cohort study. Tertiary academic center. Adult patients implanted with V1 series devices. Device integrity and speech perception testing. consonant-nucleus-consonant and AzBio in quiet speech recognition scores. At our institution, 116 V1 series cochlear implants were placed in 114 patients. Thirteen devices in prelingual patients were excluded, leaving 103 (89%) for final analysis. Forty-eight (46.6%) devices were considered as failed using the company provided EFI analysis tool. There were 36 (65.5%) of the remaining 55 devices that consistently tested within normal range; the remainder lost to follow-up with unknown status. Among the 48 device failures, 29 were revised and 19 patients were not revised. Among those not revised, 11 self-opted for observation (57.9%). Observed patients, despite impedance changes meeting failure criteria, had no subjective or objective changes in speech perception. Sentence testing scores for failure patients who elected observation (82.9 ± 11.4%) were significantly higher at failure compared with those opting for revision (55 ± 22.8%, p = 0.006). For those undergoing revision surgery, significant improvement in post-activation scores was noted as compared with time of failure with a mean improvement of 12.9% (p = 0.002, n = 24) for consonant-nucleus-consonant word scores and 17.2% (p = 0.001, n = 19) for AzBio in quiet scores. Proactive monitoring using EFI identifies a higher rate of Ultra Series V1 device failure than previously reported. However, about 20% of these patients may not have subjective change in hearing or objective decline in test scores and could be observed. Should performance worsen, reimplantation provides significant improvement in speech recognition.

Evaluating benefits of remote microphone technology for adults with hearing loss using behavioural and predictive metrics

Objective To investigate the benefit of remote-microphone (RM) systems for adults with sensory hearing loss. Design Speech recognition in quiet and in background noise was assessed. Participants with hearing loss underwent testing in two device conditions: hearing aids (HAs) alone and HAs with a RM. Normal hearing participants completed testing in the unaided condition. Predictive speech intelligibility modelling using the Hearing-Aid Speech Perception Index (HASPI) was also performed on recordings of HA processed test material. Study Sample Twenty adults with sensory hearing loss and 10 adults with normal hearing participated. Results Speech recognition for participants with hearing loss improved significantly when using the RM compared to HAs alone fit to Phonak’s proprietary prescription. Largest benefits were observed in the most challenging conditions. At the lowest signal-to-noise ratio, participants with hearing loss using a RM outperformed normal hearing listeners. Predicted intelligibility scores produced by HASPI were strongly correlated to behavioural results. Conclusions Adults using HAs who have significant difficulties understanding speech in noise will experience considerable benefits with the addition of a RM. Improvements in speech recognition were observed for all participants using RM systems, including those with relatively mild hearing loss. HASPI modelling reliably predicted the speech perception difficulties experienced.

Experienced Adult Cochlear Implant Users Show Improved Speech Recognition When Target Fitting Parameters Are Applied.

The aim of the present study was to investigate whether prediction models built by de Graaff et al. (2020) can be used to improve speech recognition in experienced adult postlingual implanted Cochlear CI users. de Graaff et al. (2020) found relationships between elevated aided thresholds and a not optimal electrical dynamic range (<50 CL or >60 CL), and poorer speech recognition in quiet and in noise. The primary hypothesis of the present study was that speech recognition improves both in quiet and in noise when the sound processor is refitted to match targets derived from the prediction models from de Graaff et al. (2020). A second hypothesis was that subjectively, most of the CI users would find the new setting too loud because of an increase in C levels, and therefore, prefer the old settings. A within-participant repeated measures design with 18 adult Cochlear CI users was used. T- and C-levels were changed to "optimized settings," as predicted by the model of de Graaff et al. (2020). Aided thresholds, speech recognition in quiet, and speech recognition in noise were measured with the old settings and after a 4-week acclimatization period with the optimized settings. Subjective benefit was measured using the Device Oriented Subjective Outcome Scale questionnaire. The mean electrical dynamic range changed from 41.1 (SD = 6.6) CL to 48.6 (SD = 3.0) CL. No significant change in aided thresholds was measured. Speech recognition improved for 16 out of 18 participants and remained stable for 2 participants. Average speech recognition scores in quiet significantly improved by 4.9% (SD = 3.8%). No significant change for speech recognition in noise was found. A significant improvement in subjective benefit was found for one of the Device Oriented Subjective Outcome subscales (speech cues) between the old and optimized settings. All participants chose to keep the optimized settings at the end of the study. We were able to improve speech recognition in quiet by optimizing the electrical dynamic range of experienced adult CI users, according to the prediction models built by de Graaff et al. (2020). There was no significant change in aided thresholds nor in speech recognition in noise. The findings of the present study suggest that improved performance for speech recognition in quiet in adult Cochlear CI users can be achieved by setting the dynamic range as close as possible to values between 50 and 60 CL when the volume level is at 10.

Multi-view Self-supervised Learning and Multi-scale Feature Fusion for Automatic Speech Recognition

To address the challenges of the poor representation capability and low data utilization rate of end-to-end speech recognition models in deep learning, this study proposes an end-to-end speech recognition model based on multi-scale feature fusion and multi-view self-supervised learning (MM-ASR). It adopts a multi-task learning paradigm for training. The proposed method emphasizes the importance of inter-layer information within shared encoders, aiming to enhance the model’s characterization capability via the multi-scale feature fusion module. Moreover, we apply multi-view self-supervised learning to effectively exploit data information. Our approach is rigorously evaluated on the Aishell-1 dataset and further validated its effectiveness on the English corpus WSJ. The experimental results demonstrate a noteworthy 4.6%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\%$$\\end{document} reduction in character error rate, indicating significantly improved speech recognition performance . These findings showcase the effectiveness and potential of our proposed MM-ASR model for end-to-end speech recognition tasks.

Cochlear-implant listeners benefit from training with time-compressed speech, even at advanced ages.

This study evaluated whether adaptive training with time-compressed speech produces an age-dependent improvement in speech recognition in 14 adult cochlear-implant users. The protocol consisted of a pretest, 5 h of training, and a posttest using time-compressed speech and an adaptive procedure. There were significant improvements in time-compressed speech recognition at the posttest session following training (>5% in the average time-compressed speech recognition threshold) but no effects of age. These results are promising for the use of adaptive training in aural rehabilitation strategies for cochlear-implant users across the adult lifespan and possibly using speech signals, such as time-compressed speech, to train temporal processing.