Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userFully Learnable Front-End for Multi-Channel Acoustic Modeling using Semi-Supervised Learning
100
0
0.0
(
0
)
Added by
Aparna Khare
Publication date
2020
fields
Informatics Engineering
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
In this work, we investigated the teacher-student training paradigm to train a fully learnable multi-channel acoustic model for far-field automatic speech recognition (ASR). Using a large offline teacher model trained on beamformed audio, we trained a simpler multi-channel student acoustic model used in the speech recognition system. For the student, both multi-channel feature extraction layers and the higher classification layers were jointly trained using the logits from the teacher model. In our experiments, compared to a baseline model trained on about 600 hours of transcribed data, a relative word-error rate (WER) reduction of about 27.3% was achieved when using an additional 1800 hours of untranscribed data. We also investigated the benefit of pre-training the multi-channel front end to output the beamformed log-mel filter bank energies (LFBE) using L2 loss. We find that pre-training improves the word error rate by 10.7% when compared to a multi-channel model directly initialized with a beamformer and mel-filter bank coefficients for the front end. Finally, combining pre-training and teacher-student training produces a WER reduction of 31% compared to our baseline.
rate research
Read More
Recent literature has shown that a learned front end with multi-channel audio input can outperform traditional beam-forming algorithms for automatic speech recognition (ASR). In this paper, we present our study on multi-channel acoustic modeling using OPUS compression with different bitrates for the different channels. We analyze the degradation in word error rate (WER) as a function of the audio encoding bitrate and show that the WER degrades by 12.6% relative with 16kpbs as compared to uncompressed audio. We show that its always preferable to have a multi-channel audio input over a single channel audio input given limited bandwidth. Our results show that for the best WER, when one of the two channels can be encoded with a bitrate higher than 32kbps, its optimal to encode the other channel with the highest bitrate possible. For bitrates lower than that, its preferable to distribute the bitrate equally between the two channels. We further show that by training the acoustic model on mixed bitrate input, up to 50% of the degradation can be recovered using a single model.
Convolutional Neural Networks have been extensively explored in the task of automatic music tagging. The problem can be approached by using either engineered time-frequency features or raw audio as input. Modulation filter bank representations that have been actively researched as a basis for timbre perception have the potential to facilitate the extraction of perceptually salient features. We explore end-to-end learned front-ends for audio representation learning, ModNet and SincModNet, that incorporate a temporal modulation processing block. The structure is effectively analogous to a modulation filter bank, where the FIR filter center frequencies are learned in a data-driven manner. The expectation is that a perceptually motivated filter bank can provide a useful representation for identifying music features. Our experimental results provide a fully visualisable and interpretable front-end temporal modulation decomposition of raw audio. We evaluate the performance of our model against the state-of-the-art of music tagging on the MagnaTagATune dataset. We analyse the impact on performance for particular tags when time-frequency bands are subsampled by the modulation filters at a progressively reduced rate. We demonstrate that modulation filtering provides promising results for music tagging and feature representation, without using extensive musical domain knowledge in the design of this front-end.
Voice assistants, such as smart speakers, have exploded in popularity. It is currently estimated that the smart speaker adoption rate has exceeded 35% in the US adult population. Manufacturers have integrated speaker identification technology, which attempts to determine the identity of the person speaking, to provide personalized services to different members of the same family. Speaker identification can also play an important role in controlling how the smart speaker is used. For example, it is not critical to correctly identify the user when playing music. However, when reading the users email out loud, it is critical to correctly verify the speaker that making the request is the authorized user. Speaker verification systems, which authenticate the speaker identity, are therefore needed as a gatekeeper to protect against various spoofing attacks that aim to impersonate the enrolled user. This paper compares popular learnable front-ends which learn the representations of audio by joint training with downstream tasks (End-to-End). We categorize the front-ends by defining two generic architectures and then analyze the filtering stages of both types in terms of learning constraints. We propose replacing fixed filterbanks with a learnable layer that can better adapt to anti-spoofing tasks. The proposed FastAudio front-end is then tested with two popular back-ends to measure the performance on the LA track of the ASVspoof 2019 dataset. The FastAudio front-end achieves a relative improvement of 27% when compared with fixed front-ends, outperforming all other learnable front-ends on this task.
Training Automatic Speech Recognition (ASR) models under federated learning (FL) settings has attracted a lot of attention recently. However, the FL scenarios often presented in the literature are artificial and fail to capture the complexity of real FL systems. In this paper, we construct a challenging and realistic ASR federated experimental setup consisting of clients with heterogeneous data distributions using the French and Italian sets of the CommonVoice dataset, a large heterogeneous dataset containing thousands of different speakers, acoustic environments and noises. We present the first empirical study on attention-based sequence-to-sequence End-to-End (E2E) ASR model with three aggregation weighting strategies -- standard FedAvg, loss-based aggregation and a novel word error rate (WER)-based aggregation, compared in two realistic FL scenarios: cross-silo with 10 clients and cross-device with 2K and 4K clients. Our analysis on E2E ASR from heterogeneous and realistic federated acoustic models provides the foundations for future research and development of realistic FL-based ASR applications.
Large scale machine learning (ML) systems such as the Alexa automatic speech recognition (ASR) system continue to improve with increasing amounts of manually transcribed training data. Instead of scaling manual transcription to impractical levels, we utilize semi-supervised learning (SSL) to learn acoustic models (AM) from the vast firehose of untranscribed audio data. Learning an AM from 1 Million hours of audio presents unique ML and system design challenges. We present the design and evaluation of a highly scalable and resource efficient SSL system for AM. Employing the student/teacher learning paradigm, we focus on the student learning subsystem: a scalable and robust data pipeline that generates features and targets from raw audio, and an efficient model pipeline, including the distributed trainer, that builds a student model. Our evaluations show that, even without extensive hyper-parameter tuning, we obtain relative accuracy improvements in the 10 to 20$%$ range, with higher gains in noisier conditions. The end-to-end processing time of this SSL system was 12 days, and several components in this system can trivially scale linearly with more compute resources.
suggested questions
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
