An official website of the United States government
Listeners typically rely more on one aspect of the speech signal than another when categorizing speech sounds. This is known as feature weighting. We present a rate distortion theory model of feature weighting and use it to ask whether human listeners select feature weights simply by mirroring the feature reliabilities that are present in their input. We show that there is an additional component (selective attention) listeners appear to use that is not reflected by the input statistics. This suggests that an internal mechanism is at play in governing listeners' weighting of different aspects of the speech signal, in addition to tracking statistics.
more »
« less
A neural architecture for selective attention to speech features
Speech perception is complex and demands constant adaptations to the speaker and the environment (i.e. noisy speech, accent, etc.). To adapt, the listener relies on one speech feature more than another. This cognitive mechanism is called selective attention. We present a model that captures the idea of selective attention: we show that this dynamic adaptation process can be captured in a neural architecture by using a multiple encoder beta variational auto encoder (beta-ME-VAE), which is based on rate distortion theory. This model implements the idea that optimal feature weighting looks different under different listening conditions and provides insight into how listeners can adapt their listening strategy on a moment-to-moment basis, even in listening situations they haven't experienced before.
more »
« less
- Award ID(s):
- 2120834
- PAR ID:
- 10425827
- Date Published:
- Journal Name:
- Proceedings of Interspeech
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Recently, speech foundation models have gained popularity due to their superiority in finetuning downstream ASR tasks. However, models finetuned on certain domains, such as LibriSpeech (adult read speech), behave poorly on other domains (child or noisy speech). One solution could be collecting as much labeled and diverse data as possible for joint finetuning on various domains. However, collecting target domain speech-text paired data and retraining the model is often costly and computationally expensive. In this paper, we introduce a simple yet effective method, speech only adaptation (SOA), based on speech foundation models (Wav2vec 2.0), which requires only speech input data from the target domain. Specifically, the Wav2vec 2.0 feature encoder is continually pretrained with the Wav2vec 2.0 loss on both the source and target domain data for domain adaptation, while the contextual encoder is frozen. Compared to a source domain finetuned model with the feature encoder being frozen during training, we find that replacing the frozen feature encoder with the adapted one provides significant WER improvements to the target domain while preserving the performance of the source domain. The effectiveness of SOA is examined on various low resource or domain mismatched ASR settings, including adult-child and clean-noisy speech.more » « less
-
Listening to speech in noise can require substantial mental effort, even among younger normal-hearing adults. The task-evoked pupil response (TEPR) has been shown to track the increased effort exerted to recognize words or sentences in increasing noise. However, few studies have examined the trajectory of listening effort across longer, more natural, stretches of speech, or the extent to which expectations about upcoming listening difficulty modulate the TEPR. Seventeen younger normal-hearing adults listened to 60-s-long audiobook passages, repeated three times in a row, at two different signal-to-noise ratios (SNRs) while pupil size was recorded. There was a significant interaction between SNR, repetition, and baseline pupil size on sustained listening effort. At lower baseline pupil sizes, potentially reflecting lower attention mobilization, TEPRs were more sustained in the harder SNR condition, particularly when attention mobilization remained low by the third presentation. At intermediate baseline pupil sizes, differences between conditions were largely absent, suggesting these listeners had optimally mobilized their attention for both SNRs. Lastly, at higher baseline pupil sizes, potentially reflecting over-mobilization of attention, the effect of SNR was initially reversed for the second and third presentations: participants initially appeared to disengage in the harder SNR condition, resulting in reduced TEPRs that recovered in the second half of the story. Together, these findings suggest that the unfolding of listening effort over time depends critically on the extent to which individuals have successfully mobilized their attention in anticipation of difficult listening conditions.more » « less
-
Non-autoregressive automatic speech recognition (NASR) models have gained attention due to their parallelism and fast inference. The encoder-based NASR, e.g. connectionist temporal classification (CTC), can be initialized from the speech foundation models (SFM) but does not account for any dependencies among intermediate tokens. The encoder-decoder-based NASR, like CTC alignment-based single-step non-autoregressive transformer (CASS-NAT), can mitigate the dependency problem but is not able to efficiently integrate SFM. Inspired by the success of recent work of speech-text joint pre-training with a shared transformer encoder, we propose a new encoder-based NASR, UniEnc-CASSNAT, to combine the advantages of CTC and CASS-NAT. UniEnc-CASSNAT consists of only an encoder as the major module, which can be the SFM. The encoder plays the role of both the CASS-NAT encoder and decoder by two forward passes. The first pass of the encoder accepts the speech signal as input, while the concatenation of the speech signal and the token-level acoustic embedding is used as the input for the second pass. Examined on the Librispeech 100 h, MyST, and Aishell1 datasets, the proposed UniEnc-CASSNAT achieves state-of-the-art NASR results and is better or comparable to CASS-NAT with only an encoder and hence, fewer model parameters.more » « less
-
While Speech Foundation Models (SFMs) excel in various speech tasks, their performance for low-resource tasks such as child Automatic Speech Recognition (ASR) is hampered by limited pretraining data. To address this, we explore different model merging techniques to leverage knowledge from models trained on larger, more diverse speech corpora. This paper also introduces Selective Attention (SA) Merge, a novel method that selectively merges task vectors from attention matrices to enhance SFM performance on low-resource tasks. Experiments on the MyST database show significant reductions in relative word error rate of up to 14%, outperforming existing model merging and data augmentation techniques. By combining data augmentation techniques with SA Merge, we achieve a new state-of-the-art WER of 8.69 on the MyST database for the Whisper-small model, highlighting the potential of SA Merge for improving low-resource ASR.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
