Search

Obtaining word timestamp information from end-to-end (E2E) ASR models remains challenging due to the lack of explicit time alignment during training. This issue is further complicated in multilingual models. Existing methods, either rely on lexicons or introduce additional tokens, leading to scalability issues and increased computational costs. In this work, we propose a new approach to estimate word boundaries without relying on lexicons. Our method leverages word embeddings from sub-word token units and a pretrained ASR model, requiring only word alignment information during training. Our proposed method can scale-up to any number of languages without incurring any additional cost. We validate our approach using a multilingual ASR model trained on five languages and demonstrate its effectiveness against a strong baseline., Comment: Submitted to ICASSP 2025

Neural audio codec tokens serve as the fundamental building blocks for speech language model (SLM)-based speech generation. However, there is no systematic understanding on how the codec system affects the speech generation performance of the SLM. In this work, we examine codec tokens within SLM framework for speech generation to provide insights for effective codec design. We retrain existing high-performing neural codec models on the same data set and loss functions to compare their performance in a uniform setting. We integrate codec tokens into two SLM systems: masked-based parallel speech generation system and an auto-regressive (AR) plus non-auto-regressive (NAR) model-based system. Our findings indicate that better speech reconstruction in codec systems does not guarantee improved speech generation in SLM. A high-quality codec decoder is crucial for natural speech production in SLM, while speech intelligibility depends more on quantization mechanism., Comment: Accepted by SLT-2024

People change their tones of voice, often accompanied by nonverbal vocalizations (NVs) such as laughter and cries, to convey rich emotions. However, most text-to-speech (TTS) systems lack the capability to generate speech with rich emotions, including NVs. This paper introduces EmoCtrl-TTS, an emotion-controllable zero-shot TTS that can generate highly emotional speech with NVs for any speaker. EmoCtrl-TTS leverages arousal and valence values, as well as laughter embeddings, to condition the flow-matching-based zero-shot TTS. To achieve high-quality emotional speech generation, EmoCtrl-TTS is trained using more than 27,000 hours of expressive data curated based on pseudo-labeling. Comprehensive evaluations demonstrate that EmoCtrl-TTS excels in mimicking the emotions of audio prompts in speech-to-speech translation scenarios. We also show that EmoCtrl-TTS can capture emotion changes, express strong emotions, and generate various NVs in zero-shot TTS. See https://aka.ms/emoctrl-tts for demo samples., Comment: Accepted by SLT2024. See https://aka.ms/emoctrl-tts for demo samples

We present MELLE, a novel continuous-valued tokens based language modeling approach for text to speech synthesis (TTS). MELLE autoregressively generates continuous mel-spectrogram frames directly from text condition, bypassing the need for vector quantization, which are originally designed for audio compression and sacrifice fidelity compared to mel-spectrograms. Specifically, (i) instead of cross-entropy loss, we apply regression loss with a proposed spectrogram flux loss function to model the probability distribution of the continuous-valued tokens. (ii) we have incorporated variational inference into MELLE to facilitate sampling mechanisms, thereby enhancing the output diversity and model robustness. Experiments demonstrate that, compared to the two-stage codec language models VALL-E and its variants, the single-stage MELLE mitigates robustness issues by avoiding the inherent flaws of sampling discrete codes, achieves superior performance across multiple metrics, and, most importantly, offers a more streamlined paradigm. See https://aka.ms/melle for demos of our work.

With the help of discrete neural audio codecs, large language models (LLM) have increasingly been recognized as a promising methodology for zero-shot Text-to-Speech (TTS) synthesis. However, sampling based decoding strategies bring astonishing diversity to generation, but also pose robustness issues such as typos, omissions and repetition. In addition, the high sampling rate of audio also brings huge computational overhead to the inference process of autoregression. To address these issues, we propose VALL-E R, a robust and efficient zero-shot TTS system, building upon the foundation of VALL-E. Specifically, we introduce a phoneme monotonic alignment strategy to strengthen the connection between phonemes and acoustic sequence, ensuring a more precise alignment by constraining the acoustic tokens to match their associated phonemes. Furthermore, we employ a codec-merging approach to downsample the discrete codes in shallow quantization layer, thereby accelerating the decoding speed while preserving the high quality of speech output. Benefiting from these strategies, VALL-E R obtains controllablity over phonemes and demonstrates its strong robustness by approaching the WER of ground truth. In addition, it requires fewer autoregressive steps, with over 60% time reduction during inference. This research has the potential to be applied to meaningful projects, including the creation of speech for those affected by aphasia. Audio samples will be available at: https://aka.ms/valler., Comment: 15 pages, 5 figures

Language-agnostic many-to-one end-to-end speech translation models can convert audio signals from different source languages into text in a target language. These models do not need source language identification, which improves user experience. In some cases, the input language can be given or estimated. Our goal is to use this additional language information while preserving the quality of the other languages. We accomplish this by introducing a simple and effective linear input network. The linear input network is initialized as an identity matrix, which ensures that the model can perform as well as, or better than, the original model. Experimental results show that the proposed method can successfully enhance the specified language, while keeping the language-agnostic ability of the many-to-one ST models.

Recently, zero-shot text-to-speech (TTS) systems, capable of synthesizing any speaker's voice from a short audio prompt, have made rapid advancements. However, the quality of the generated speech significantly deteriorates when the audio prompt contains noise, and limited research has been conducted to address this issue. In this paper, we explored various strategies to enhance the quality of audio generated from noisy audio prompts within the context of flow-matching-based zero-shot TTS. Our investigation includes comprehensive training strategies: unsupervised pre-training with masked speech denoising, multi-speaker detection and DNSMOS-based data filtering on the pre-training data, and fine-tuning with random noise mixing. The results of our experiments demonstrate significant improvements in intelligibility, speaker similarity, and overall audio quality compared to the approach of applying speech enhancement to the audio prompt., Comment: Accepted to INTERSPEECH2024

This paper introduces VALL-E 2, the latest advancement in neural codec language models that marks a milestone in zero-shot text-to-speech synthesis (TTS), achieving human parity for the first time. Based on its predecessor, VALL-E, the new iteration introduces two significant enhancements: Repetition Aware Sampling refines the original nucleus sampling process by accounting for token repetition in the decoding history. It not only stabilizes the decoding but also circumvents the infinite loop issue. Grouped Code Modeling organizes codec codes into groups to effectively shorten the sequence length, which not only boosts inference speed but also addresses the challenges of long sequence modeling. Our experiments on the LibriSpeech and VCTK datasets show that VALL-E 2 surpasses previous systems in speech robustness, naturalness, and speaker similarity. It is the first of its kind to reach human parity on these benchmarks. Moreover, VALL-E 2 consistently synthesizes high-quality speech, even for sentences that are traditionally challenging due to their complexity or repetitive phrases. The advantages of this work could contribute to valuable endeavors, such as generating speech for individuals with aphasia or people with amyotrophic lateral sclerosis. See https://aka.ms/valle2 for demos of VALL-E 2., Comment: Demo posted

Accurate control of the total duration of generated speech by adjusting the speech rate is crucial for various text-to-speech (TTS) applications. However, the impact of adjusting the speech rate on speech quality, such as intelligibility and speaker characteristics, has been underexplored. In this work, we propose a novel total-duration-aware (TDA) duration model for TTS, where phoneme durations are predicted not only from the text input but also from an additional input of the total target duration. We also propose a MaskGIT-based duration model that enhances the diversity and quality of the predicted phoneme durations. Our results demonstrate that the proposed TDA duration models achieve better intelligibility and speaker similarity for various speech rate configurations compared to the baseline models. We also show that the proposed MaskGIT-based model can generate phoneme durations with higher quality and diversity compared to its regression or flow-matching counterparts., Comment: Accepted to Interspeech 2024

There is a rising interest and trend in research towards directly translating speech from one language to another, known as end-to-end speech-to-speech translation. However, most end-to-end models struggle to outperform cascade models, i.e., a pipeline framework by concatenating speech recognition, machine translation and text-to-speech models. The primary challenges stem from the inherent complexities involved in direct translation tasks and the scarcity of data. In this study, we introduce a novel model framework TransVIP that leverages diverse datasets in a cascade fashion yet facilitates end-to-end inference through joint probability. Furthermore, we propose two separated encoders to preserve the speaker's voice characteristics and isochrony from the source speech during the translation process, making it highly suitable for scenarios such as video dubbing. Our experiments on the French-English language pair demonstrate that our model outperforms the current state-of-the-art speech-to-speech translation model., Comment: Work in progress

Recent advancements in zero-shot text-to-speech (TTS) modeling have led to significant strides in generating high-fidelity and diverse speech. However, dialogue generation, along with achieving human-like naturalness in speech, continues to be a challenge. In this paper, we introduce CoVoMix: Conversational Voice Mixture Generation, a novel model for zero-shot, human-like, multi-speaker, multi-round dialogue speech generation. CoVoMix first converts dialogue text into multiple streams of discrete tokens, with each token stream representing semantic information for individual talkers. These token streams are then fed into a flow-matching based acoustic model to generate mixed mel-spectrograms. Finally, the speech waveforms are produced using a HiFi-GAN model. Furthermore, we devise a comprehensive set of metrics for measuring the effectiveness of dialogue modeling and generation. Our experimental results show that CoVoMix can generate dialogues that are not only human-like in their naturalness and coherence but also involve multiple talkers engaging in multiple rounds of conversation. This is exemplified by instances generated in a single channel where one speaker's utterance is seamlessly mixed with another's interjections or laughter, indicating the latter's role as an attentive listener. Audio samples are available at https://aka.ms/covomix.

We present RALL-E, a robust language modeling method for text-to-speech (TTS) synthesis. While previous work based on large language models (LLMs) shows impressive performance on zero-shot TTS, such methods often suffer from poor robustness, such as unstable prosody (weird pitch and rhythm/duration) and a high word error rate (WER), due to the autoregressive prediction style of language models. The core idea behind RALL-E is chain-of-thought (CoT) prompting, which decomposes the task into simpler steps to enhance the robustness of LLM-based TTS. To accomplish this idea, RALL-E first predicts prosody features (pitch and duration) of the input text and uses them as intermediate conditions to predict speech tokens in a CoT style. Second, RALL-E utilizes the predicted duration prompt to guide the computing of self-attention weights in Transformer to enforce the model to focus on the corresponding phonemes and prosody features when predicting speech tokens. Results of comprehensive objective and subjective evaluations demonstrate that, compared to a powerful baseline method VALL-E, RALL-E significantly improves the WER of zero-shot TTS from $5.6\%$ (without reranking) and $1.7\%$ (with reranking) to $2.5\%$ and $1.0\%$, respectively. Furthermore, we demonstrate that RALL-E correctly synthesizes sentences that are hard for VALL-E and reduces the error rate from $68\%$ to $4\%$.

The recent advancements in large language models (LLMs) have revolutionized the field of natural language processing, progressively broadening their scope to multimodal perception and generation. However, effectively integrating listening capabilities into LLMs poses significant challenges, particularly with respect to generalizing across varied contexts and executing complex auditory tasks. In this work, we introduce WavLLM, a robust and adaptive speech large language model with dual encoders, and a prompt-aware LoRA weight adapter, optimized by a two-stage curriculum learning approach. Leveraging dual encoders, we decouple different types of speech information, utilizing a Whisper encoder to process the semantic content of speech, and a WavLM encoder to capture the unique characteristics of the speaker's identity. Within the curriculum learning framework, WavLLM first builds its foundational capabilities by optimizing on mixed elementary single tasks, followed by advanced multi-task training on more complex tasks such as combinations of the elementary tasks. To enhance the flexibility and adherence to different tasks and instructions, a prompt-aware LoRA weight adapter is introduced in the second advanced multi-task training stage. We validate the proposed model on universal speech benchmarks including tasks such as ASR, ST, SV, ER, and also apply it to specialized datasets like Gaokao English listening comprehension set for SQA, and speech Chain-of-Thought (CoT) evaluation set. Experiments demonstrate that the proposed model achieves state-of-the-art performance across a range of speech tasks on the same model size, exhibiting robust generalization capabilities in executing complex tasks using CoT approach. Furthermore, our model successfully completes Gaokao tasks without specialized training. The codes, models, audio, and Gaokao evaluation set can be accessed at \url{aka.ms/wavllm}., Comment: accepted by EMNLP2024 findings

In this paper, we propose two novel approaches, which integrate long-content information into the factorized neural transducer (FNT) based architecture in both non-streaming (referred to as LongFNT ) and streaming (referred to as SLongFNT ) scenarios. We first investigate whether long-content transcriptions can improve the vanilla conformer transducer (C-T) models. Our experiments indicate that the vanilla C-T models do not exhibit improved performance when utilizing long-content transcriptions, possibly due to the predictor network of C-T models not functioning as a pure language model. Instead, FNT shows its potential in utilizing long-content information, where we propose the LongFNT model and explore the impact of long-content information in both text (LongFNT-Text) and speech (LongFNT-Speech). The proposed LongFNT-Text and LongFNT-Speech models further complement each other to achieve better performance, with transcription history proving more valuable to the model. The effectiveness of our LongFNT approach is evaluated on LibriSpeech and GigaSpeech corpora, and obtains relative 19% and 12% word error rate reduction, respectively. Furthermore, we extend the LongFNT model to the streaming scenario, which is named SLongFNT , consisting of SLongFNT-Text and SLongFNT-Speech approaches to utilize long-content text and speech information. Experiments show that the proposed SLongFNT model achieves relative 26% and 17% WER reduction on LibriSpeech and GigaSpeech respectively while keeping a good latency, compared to the FNT baseline. Overall, our proposed LongFNT and SLongFNT highlight the significance of considering long-content speech and transcription knowledge for improving both non-streaming and streaming speech recognition systems., Comment: Accepted by TASLP 2024

While recent large-scale text-to-speech (TTS) models have achieved significant progress, they still fall short in speech quality, similarity, and prosody. Considering speech intricately encompasses various attributes (e.g., content, prosody, timbre, and acoustic details) that pose significant challenges for generation, a natural idea is to factorize speech into individual subspaces representing different attributes and generate them individually. Motivated by it, we propose NaturalSpeech 3, a TTS system with novel factorized diffusion models to generate natural speech in a zero-shot way. Specifically, 1) we design a neural codec with factorized vector quantization (FVQ) to disentangle speech waveform into subspaces of content, prosody, timbre, and acoustic details; 2) we propose a factorized diffusion model to generate attributes in each subspace following its corresponding prompt. With this factorization design, NaturalSpeech 3 can effectively and efficiently model intricate speech with disentangled subspaces in a divide-and-conquer way. Experiments show that NaturalSpeech 3 outperforms the state-of-the-art TTS systems on quality, similarity, prosody, and intelligibility, and achieves on-par quality with human recordings. Furthermore, we achieve better performance by scaling to 1B parameters and 200K hours of training data., Comment: Achieving human-level quality and naturalness on multi-speaker datasets (e.g., LibriSpeech) in a zero-shot way

Large language models (LLMs) have made significant advancements in natural language processing and are concurrently extending the language ability to other modalities, such as speech and vision. Nevertheless, most of the previous work focuses on prompting LLMs with perception abilities like auditory comprehension, and the effective approach for augmenting LLMs with speech synthesis capabilities remains ambiguous. In this paper, we conduct a comprehensive empirical exploration of boosting LLMs with the ability to generate speech, by combining pre-trained LLM LLaMA/OPT and text-to-speech synthesis model VALL-E. We compare three integration methods between LLMs and speech synthesis models, including directly fine-tuned LLMs, superposed layers of LLMs and VALL-E, and coupled LLMs and VALL-E using LLMs as a powerful text encoder. Experimental results show that, using LoRA method to fine-tune LLMs directly to boost the speech synthesis capability does not work well, and superposed LLMs and VALL-E can improve the quality of generated speech both in speaker similarity and word error rate (WER). Among these three methods, coupled methods leveraging LLMs as the text encoder can achieve the best performance, making it outperform original speech synthesis models with a consistently better speaker similarity and a significant (10.9%) WER reduction.

The development of U.S. Army and NATO data link systems is introduced first, and then the development trend of future intelligent data link is summarized into integration, generalization, multifunctionality and high security. A unit-level combat system architecture based on the global combat cloud, which is capable of realizing the flexible scheduling of global combat resources and maximizing the overall combat effectiveness, is proposed. Intelligent data link is an important part of this solution, providing strong information support for future urban unit-level warfare.

We present a cost-effective method to integrate speech into a large language model (LLM), resulting in a Contextual Speech Model with Instruction-following/in-context-learning Capabilities (COSMIC) multi-modal LLM. Using GPT-3.5, we generate Speech Comprehension Test Question-Answer (SQA) pairs from speech transcriptions for supervised instruction tuning. With under 30 million trainable parameters and only 450 hours of English speech data, COSMIC demonstrates emerging capabilities in instruction-following and in-context learning. Equipped with such capabilities, COSMIC achieves a maximum 33.18 BLEU score in 0-shot EN-to-X speech to text translation (S2TT) and a significant boost in the 1-shot setting. Additionally, there is an average 25.8\% relative Word Error Rate (WER) reduction for 1-shot cross-domain adaptation. COSMIC exhibits a significant automatic speech recognition (ASR) accuracy gain in contextual biasing tasks due to its instruction-following capability.

The growing need for instant spoken language transcription and translation is driven by increased global communication and cross-lingual interactions. This has made offering translations in multiple languages essential for user applications. Traditional approaches to automatic speech recognition (ASR) and speech translation (ST) have often relied on separate systems, leading to inefficiencies in computational resources, and increased synchronization complexity in real time. In this paper, we propose a streaming Transformer-Transducer (T-T) model able to jointly produce many-to-one and one-to-many transcription and translation using a single decoder. We introduce a novel method for joint token-level serialized output training based on timestamp information to effectively produce ASR and ST outputs in the streaming setting. Experiments on {it,es,de}->en prove the effectiveness of our approach, enabling the generation of one-to-many joint outputs with a single decoder for the first time., Comment: \c{opyright} 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works

It is typically challenging for visual or visual-inertial odometry systems to handle the problems of dynamic scenes and pure rotation. In this work, we design a novel visual-inertial odometry (VIO) system called RD-VIO to handle both of these two problems. Firstly, we propose an IMU-PARSAC algorithm which can robustly detect and match keypoints in a two-stage process. In the first state, landmarks are matched with new keypoints using visual and IMU measurements. We collect statistical information from the matching and then guide the intra-keypoint matching in the second stage. Secondly, to handle the problem of pure rotation, we detect the motion type and adapt the deferred-triangulation technique during the data-association process. We make the pure-rotational frames into the special subframes. When solving the visual-inertial bundle adjustment, they provide additional constraints to the pure-rotational motion. We evaluate the proposed VIO system on public datasets and online comparison. Experiments show the proposed RD-VIO has obvious advantages over other methods in dynamic environments. The source code is available at: \href{https://github.com/openxrlab/xrslam}{{\fontfamily{pcr}\selectfont https://github.com/openxrlab/xrslam}}.

Due to the powerful edge-preserving ability and low computational complexity, Guided image filter (GIF) and its improved versions has been widely applied in computer vision and image processing. However, all of them are suffered halo artifacts to some degree, as the regularization parameter increase. In the case of inconsistent structure of guidance image and input image, edge-preserving ability degradation will also happen. In this paper, a novel guided image filter is proposed by integrating an explicit first-order edge-protect constraint and an explicit residual constraint which will improve the edge-preserving ability in both cases. To illustrate the efficiency of the proposed filter, the performances are shown in some typical applications, which are single image detail enhancement, multi-scale exposure fusion, hyper spectral images classification. Both theoretical analysis and experimental results prove that the powerful edge-preserving ability of the proposed filter.

Simultaneous Speech-to-Text translation serves a critical role in real-time crosslingual communication. Despite the advancements in recent years, challenges remain in achieving stability in the translation process, a concern primarily manifested in the flickering of partial results. In this paper, we propose a novel revision-controllable method designed to address this issue. Our method introduces an allowed revision window within the beam search pruning process to screen out candidate translations likely to cause extensive revisions, leading to a substantial reduction in flickering and, crucially, providing the capability to completely eliminate flickering. The experiments demonstrate the proposed method can significantly improve the decoding stability without compromising substantially on the translation quality., Comment: accepted by ASRU 2023

Memory constraint of always-on devices is one of the major concerns when deploying speech processing models on these devices. While larger models trained with sufficiently large amount of data generally perform better, making them fit in the device memory is a demanding challenge. In this paper, we aim to reduce model size by reparameterizing model weights across Transformer encoder layers and assuming a special weight composition and structure. More specifically, inspired by ResNet and the more recent LoRA work, we propose an approach named ResidualTransformer, where each weight matrix in a Transformer layer comprises 1) a shared full-rank component with its adjacent layers, and 2) a unique low-rank component to itself. The low-rank matrices only account for a small amount of model size increase. In addition, we add diagonal weight matrices to improve modeling capacity of the low-rank matrices. Experiments of our 10k-hour speech recognition and speech translation tasks show that the Transformer encoder size can be reduced by ~3X with very slight performance degradation., Comment: Accepted at IEEE ICASSP 2024. 5 pages, 1 figure

Token-level serialized output training (t-SOT) was recently proposed to address the challenge of streaming multi-talker automatic speech recognition (ASR). T-SOT effectively handles overlapped speech by representing multi-talker transcriptions as a single token stream with $\langle \text{cc}\rangle$ symbols interspersed. However, the use of a naive neural transducer architecture significantly constrained its applicability for text-only adaptation. To overcome this limitation, we propose a novel t-SOT model structure that incorporates the idea of factorized neural transducers (FNT). The proposed method separates a language model (LM) from the transducer's predictor and handles the unnatural token order resulting from the use of $\langle \text{cc}\rangle$ symbols in t-SOT. We achieve this by maintaining multiple hidden states and introducing special handling of the $\langle \text{cc}\rangle$ tokens within the LM. The proposed t-SOT FNT model achieves comparable performance to the original t-SOT model while retaining the ability to reduce word error rate (WER) on both single and multi-talker datasets through text-only adaptation., Comment: 5 pages, 2 figures, submitted to ICASSP2024

End-to-end speech translation (ST) for conversation recordings involves several under-explored challenges such as speaker diarization (SD) without accurate word time stamps and handling of overlapping speech in a streaming fashion. In this work, we propose DiariST, the first streaming ST and SD solution. It is built upon a neural transducer-based streaming ST system and integrates token-level serialized output training and t-vector, which were originally developed for multi-talker speech recognition. Due to the absence of evaluation benchmarks in this area, we develop a new evaluation dataset, DiariST-AliMeeting, by translating the reference Chinese transcriptions of the AliMeeting corpus into English. We also propose new metrics, called speaker-agnostic BLEU and speaker-attributed BLEU, to measure the ST quality while taking SD accuracy into account. Our system achieves a strong ST and SD capability compared to offline systems based on Whisper, while performing streaming inference for overlapping speech. To facilitate the research in this new direction, we release the evaluation data, the offline baseline systems, and the evaluation code., Comment: Accepted to ICASSP 2024

Recent advancements in generative speech models based on audio-text prompts have enabled remarkable innovations like high-quality zero-shot text-to-speech. However, existing models still face limitations in handling diverse audio-text speech generation tasks involving transforming input speech and processing audio captured in adverse acoustic conditions. This paper introduces SpeechX, a versatile speech generation model capable of zero-shot TTS and various speech transformation tasks, dealing with both clean and noisy signals. SpeechX combines neural codec language modeling with multi-task learning using task-dependent prompting, enabling unified and extensible modeling and providing a consistent way for leveraging textual input in speech enhancement and transformation tasks. Experimental results show SpeechX's efficacy in various tasks, including zero-shot TTS, noise suppression, target speaker extraction, speech removal, and speech editing with or without background noise, achieving comparable or superior performance to specialized models across tasks. See https://aka.ms/speechx for demo samples., Comment: To appear in TASLP. See https://aka.ms/speechx for demo samples

Metaverse has attracted great attention from industry and academia in recent years. Metaverse for the ocean (Meta-ocean) is the implementation of the Metaverse technologies in virtual emersion of the ocean which is beneficial for people yearning for the ocean. It has demonstrated great potential for tourism and education with its strong immersion and appealing interactive user experience. However, quite limited endeavors have been spent on exploring the full possibility of Meta-ocean, especially in modeling the movements of marine creatures. In this paper, we first investigate the technology status of Metaverse and virtual reality (VR) and develop a prototype that builds the Meta-ocean in VR devices with strong immersive visual effects. Then, we demonstrate a method to model the undersea scene and marine creatures and propose an optimized path algorithm based on the Catmull-Rom spline to model the movements of marine life. Finally, we conduct a user study to analyze our Meta-ocean prototype. This user study illustrates that our new prototype can give us strong immersion and an appealing interactive user experience.

In end-to-end automatic speech recognition system, one of the difficulties for language expansion is the limited paired speech and text training data. In this paper, we propose a novel method to generate augmented samples with unpaired speech feature segments and text data for model pre-training, which has the advantage of low cost without using additional speech data. When mixing 20,000 hours augmented speech data generated by our method with 12,500 hours original transcribed speech data for Italian Transformer transducer model pre-training, we achieve 8.7% relative word error rate reduction. The pre-trained model achieves similar performance as the model pre-trained with multilingual transcribed 75,000 hours raw speech data. When merging the augmented speech data with the multilingual data to pre-train a new model, we achieve even more relative word error rate reduction of 12.2% over the baseline, which further verifies the effectiveness of our method for speech data augmentation.

Large language models (LLMs) have achieved remarkable success in the field of natural language processing, enabling better human-computer interaction using natural language. However, the seamless integration of speech signals into LLMs has not been explored well. The "decoder-only" architecture has also not been well studied for speech processing tasks. In this research, we introduce Speech-LLaMA, a novel approach that effectively incorporates acoustic information into text-based large language models. Our method leverages Connectionist Temporal Classification and a simple audio encoder to map the compressed acoustic features to the continuous semantic space of the LLM. In addition, we further probe the decoder-only architecture for speech-to-text tasks by training a smaller scale randomly initialized speech-LLaMA model from speech-text paired data alone. We conduct experiments on multilingual speech-to-text translation tasks and demonstrate a significant improvement over strong baselines, highlighting the potential advantages of decoder-only models for speech-to-text conversion.

In real-world applications, users often require both translations and transcriptions of speech to enhance their comprehension, particularly in streaming scenarios where incremental generation is necessary. This paper introduces a streaming Transformer-Transducer that jointly generates automatic speech recognition (ASR) and speech translation (ST) outputs using a single decoder. To produce ASR and ST content effectively with minimal latency, we propose a joint token-level serialized output training method that interleaves source and target words by leveraging an off-the-shelf textual aligner. Experiments in monolingual (it-en) and multilingual (\{de,es,it\}-en) settings demonstrate that our approach achieves the best quality-latency balance. With an average ASR latency of 1s and ST latency of 1.3s, our model shows no degradation or even improves output quality compared to separate ASR and ST models, yielding an average improvement of 1.1 WER and 0.4 BLEU in the multilingual case., Comment: Accepted at ASRU 2023

Recent end-to-end automatic speech recognition (ASR) systems often utilize a Transformer-based acoustic encoder that generates embedding at a high frame rate. However, this design is inefficient, particularly for long speech signals due to the quadratic computation of self-attention. To address this, we propose a new method, Adjacent Token Merging (A-ToMe), which gradually combines adjacent tokens with high similarity scores between their key values. In this way, the total time step could be reduced, and the inference of both the encoder and joint network is accelerated. Experiments on LibriSpeech show that our method can reduce 57% of tokens and improve the inference speed on GPU by 70% without any notable loss of accuracy. Additionally, we demonstrate that A-ToMe is also an effective solution to reduce tokens in long-form ASR, where the input speech consists of multiple utterances., Comment: Interspeech 2023

The integration of Language Models (LMs) has proven to be an effective way to address domain shifts in speech recognition. However, these approaches usually require a significant amount of target domain text data for the training of LMs. Different from these methods, in this work, with only a domain-specific text prompt, we propose two zero-shot ASR domain adaptation methods using LLaMA, a 7-billion-parameter large language model (LLM). LLM is used in two ways: 1) second-pass rescoring: reranking N-best hypotheses of a given ASR system with LLaMA; 2) deep LLM-fusion: incorporating LLM into the decoder of an encoder-decoder based ASR system. Experiments show that, with only one domain prompt, both methods can effectively reduce word error rates (WER) on out-of-domain TedLium-2 and SPGISpeech datasets. Especially, the deep LLM-fusion has the advantage of better recall of entity and out-of-vocabulary words.

Automatic Speech Recognition (ASR) has seen remarkable advancements with deep neural networks, such as Transformer and Conformer. However, these models typically have large model sizes and high inference costs, posing a challenge to deploy on resource-limited devices. In this paper, we propose a novel compression strategy that leverages structured pruning and knowledge distillation to reduce the model size and inference cost of the Conformer model while preserving high recognition performance. Our approach utilizes a set of binary masks to indicate whether to retain or prune each Conformer module, and employs L0 regularization to learn the optimal mask values. To further enhance pruning performance, we use a layerwise distillation strategy to transfer knowledge from unpruned to pruned models. Our method outperforms all pruning baselines on the widely used LibriSpeech benchmark, achieving a 50% reduction in model size and a 28% reduction in inference cost with minimal performance loss., Comment: Accepted at INTERSPEECH 2023

Recent research shows a big convergence in model architecture, training objectives, and inference methods across various tasks for different modalities. In this paper, we propose VioLA, a single auto-regressive Transformer decoder-only network that unifies various cross-modal tasks involving speech and text, such as speech-to-text, text-to-text, text-to-speech, and speech-to-speech tasks, as a conditional codec language model task via multi-task learning framework. To accomplish this, we first convert all the speech utterances to discrete tokens (similar to the textual data) using an offline neural codec encoder. In such a way, all these tasks are converted to token-based sequence conversion problems, which can be naturally handled with one conditional language model. We further integrate task IDs (TID) and language IDs (LID) into the proposed model to enhance the modeling capability of handling different languages and tasks. Experimental results demonstrate that the proposed VioLA model can support both single-modal and cross-modal tasks well, and the decoder-only model achieves a comparable and even better performance than the strong baselines., Comment: Working in progress

In order to deal with the sparse and unstructured raw point clouds, LiDAR based 3D object detection research mostly focuses on designing dedicated local point aggregators for fine-grained geometrical modeling. In this paper, we revisit the local point aggregators from the perspective of allocating computational resources. We find that the simplest pillar based models perform surprisingly well considering both accuracy and latency. Additionally, we show that minimal adaptions from the success of 2D object detection, such as enlarging receptive field, significantly boost the performance. Extensive experiments reveal that our pillar based networks with modernized designs in terms of architecture and training render the state-of-the-art performance on the two popular benchmarks: Waymo Open Dataset and nuScenes. Our results challenge the common intuition that the detailed geometry modeling is essential to achieve high performance for 3D object detection., Comment: CVPR 2023

We propose a cross-lingual neural codec language model, VALL-E X, for cross-lingual speech synthesis. Specifically, we extend VALL-E and train a multi-lingual conditional codec language model to predict the acoustic token sequences of the target language speech by using both the source language speech and the target language text as prompts. VALL-E X inherits strong in-context learning capabilities and can be applied for zero-shot cross-lingual text-to-speech synthesis and zero-shot speech-to-speech translation tasks. Experimental results show that it can generate high-quality speech in the target language via just one speech utterance in the source language as a prompt while preserving the unseen speaker's voice, emotion, and acoustic environment. Moreover, VALL-E X effectively alleviates the foreign accent problems, which can be controlled by a language ID. Audio samples are available at \url{https://aka.ms/vallex}., Comment: We encourage readers to listen to the audio samples on our demo page: \url{https://aka.ms/vallex}

We propose gated language experts and curriculum training to enhance multilingual transformer transducer models without requiring language identification (LID) input from users during inference. Our method incorporates a gating mechanism and LID loss, enabling transformer experts to learn language-specific information. By combining gated transformer experts with shared transformer layers, we construct multilingual transformer blocks and utilize linear experts to effectively regularize the joint network. The curriculum training scheme leverages LID to guide the gated experts in improving their respective language performance. Experimental results on a bilingual task involving English and Spanish demonstrate significant improvements, with average relative word error reductions of 12.5% and 7.3% compared to the baseline bilingual and monolingual models, respectively. Notably, our method achieves performance comparable to the upper-bound model trained and inferred with oracle LID. Extending our approach to trilingual, quadrilingual, and pentalingual models reveals similar advantages to those observed in the bilingual models, highlighting its ease of extension to multiple languages.

We previously proposed contextual spelling correction (CSC) to correct the output of end-to-end (E2E) automatic speech recognition (ASR) models with contextual information such as name, place, etc. Although CSC has achieved reasonable improvement in the biasing problem, there are still two drawbacks for further accuracy improvement. First, due to information limitation in text only hypothesis or weak performance of ASR model on rare domains, the CSC model may fail to correct phrases with similar pronunciation or anti-context cases where all biasing phrases are not present in the utterance. Second, there is a discrepancy between the training and inference of CSC. The bias list in training is randomly selected but in inference there may be more similarity between ground truth phrase and other phrases. To solve above limitations, in this paper we propose an improved non-autoregressive (NAR) spelling correction model for contextual biasing in E2E neural transducer-based ASR systems to improve the previous CSC model from two perspectives: Firstly, we incorporate acoustics information with an external attention as well as text hypotheses into CSC to better distinguish target phrase from dissimilar or irrelevant phrases. Secondly, we design a semantic aware data augmentation schema in training phrase to reduce the mismatch between training and inference to further boost the biasing accuracy. Experiments show that the improved method outperforms the baseline ASR+Biasing system by as much as 20.3% relative name recall gain and achieves stable improvement compared to the previous CSC method over different bias list name coverage ratio., Comment: Accepted by ICASSP 2023

Speaker change detection (SCD) is an important feature that improves the readability of the recognized words from an automatic speech recognition (ASR) system by breaking the word sequence into paragraphs at speaker change points. Existing SCD solutions either require additional ensemble for the time based decisions and recognized word sequences, or implement a tight integration between ASR and SCD, limiting the potential optimum performance for both tasks. To address these issues, we propose a novel framework for the SCD task, where an additional SCD module is built on top of an existing Transformer Transducer ASR (TT-ASR) network. Two variants of the SCD network are explored in this framework that naturally estimate speaker change probability for each word, while allowing the ASR and SCD to have independent optimization scheme for the best performance. Experiments show that our methods can significantly improve the F1 score on LibriCSS and Microsoft call center data sets without ASR degradation, compared with a joint SCD and ASR baseline., Comment: 5 pages, 1 figure, accepted by ICASSP 2023

We introduce a language modeling approach for text to speech synthesis (TTS). Specifically, we train a neural codec language model (called Vall-E) using discrete codes derived from an off-the-shelf neural audio codec model, and regard TTS as a conditional language modeling task rather than continuous signal regression as in previous work. During the pre-training stage, we scale up the TTS training data to 60K hours of English speech which is hundreds of times larger than existing systems. Vall-E emerges in-context learning capabilities and can be used to synthesize high-quality personalized speech with only a 3-second enrolled recording of an unseen speaker as an acoustic prompt. Experiment results show that Vall-E significantly outperforms the state-of-the-art zero-shot TTS system in terms of speech naturalness and speaker similarity. In addition, we find Vall-E could preserve the speaker's emotion and acoustic environment of the acoustic prompt in synthesis. See https://aka.ms/valle for demos of our work., Comment: Working in progress

Neural transducer is now the most popular end-to-end model for speech recognition, due to its naturally streaming ability. However, it is challenging to adapt it with text-only data. Factorized neural transducer (FNT) model was proposed to mitigate this problem. The improved adaptation ability of FNT on text-only adaptation data came at the cost of lowered accuracy compared to the standard neural transducer model. We propose several methods to improve the performance of the FNT model. They are: adding CTC criterion during training, adding KL divergence loss during adaptation, using a pre-trained language model to seed the vocabulary predictor, and an efficient adaptation approach by interpolating the vocabulary predictor with the n-gram language model. A combination of these approaches results in a relative word-error-rate reduction of 9.48\% from the standard FNT model. Furthermore, n-gram interpolation with the vocabulary predictor improves the adaptation speed hugely with satisfactory adaptation performance.

Although speech is a simple and effective way for humans to communicate with the outside world, a more realistic speech interaction contains multimodal information, e.g., vision, text. How to design a unified framework to integrate different modal information and leverage different resources (e.g., visual-audio pairs, audio-text pairs, unlabeled speech, and unlabeled text) to facilitate speech representation learning was not well explored. In this paper, we propose a unified cross-modal representation learning framework VATLM (Visual-Audio-Text Language Model). The proposed VATLM employs a unified backbone network to model the modality-independent information and utilizes three simple modality-dependent modules to preprocess visual, speech, and text inputs. In order to integrate these three modalities into one shared semantic space, VATLM is optimized with a masked prediction task of unified tokens, given by our proposed unified tokenizer. We evaluate the pre-trained VATLM on audio-visual related downstream tasks, including audio-visual speech recognition (AVSR), visual speech recognition (VSR) tasks. Results show that the proposed VATLM outperforms previous the state-of-the-art models, such as audio-visual pre-trained AV-HuBERT model, and analysis also demonstrates that VATLM is capable of aligning different modalities into the same space. To facilitate future research, we release the code and pre-trained models at https://aka.ms/vatlm., Comment: 11 pages, Accepted by IEEE Transactions on Multimedia

Traditional automatic speech recognition~(ASR) systems usually focus on individual utterances, without considering long-form speech with useful historical information, which is more practical in real scenarios. Simply attending longer transcription history for a vanilla neural transducer model shows no much gain in our preliminary experiments, since the prediction network is not a pure language model. This motivates us to leverage the factorized neural transducer structure, containing a real language model, the vocabulary predictor. We propose the {LongFNT-Text} architecture, which fuses the sentence-level long-form features directly with the output of the vocabulary predictor and then embeds token-level long-form features inside the vocabulary predictor, with a pre-trained contextual encoder RoBERTa to further boost the performance. Moreover, we propose the {LongFNT} architecture by extending the long-form speech to the original speech input and achieve the best performance. The effectiveness of our LongFNT approach is validated on LibriSpeech and GigaSpeech corpora with 19% and 12% relative word error rate~(WER) reduction, respectively., Comment: Submitted to ICASSP2023

Self-supervised learning (SSL), which utilizes the input data itself for representation learning, has achieved state-of-the-art results for various downstream speech tasks. However, most of the previous studies focused on offline single-talker applications, with limited investigations in multi-talker cases, especially for streaming scenarios. In this paper, we investigate SSL for streaming multi-talker speech recognition, which generates transcriptions of overlapping speakers in a streaming fashion. We first observe that conventional SSL techniques do not work well on this task due to the poor representation of overlapping speech. We then propose a novel SSL training objective, referred to as bi-label masked speech prediction, which explicitly preserves representations of all speakers in overlapping speech. We investigate various aspects of the proposed system including data configuration and quantizer selection. The proposed SSL setup achieves substantially better word error rates on the LibriSpeechMix dataset., Comment: submitted to ICASSP 2023

Self-supervised learning (SSL) methods such as WavLM have shown promising speech separation (SS) results in small-scale simulation-based experiments. In this work, we extend the exploration of the SSL-based SS by massively scaling up both the pre-training data (more than 300K hours) and fine-tuning data (10K hours). We also investigate various techniques to efficiently integrate the pre-trained model with the SS network under a limited computation budget, including a low frame rate SSL model training setup and a fine-tuning scheme using only the part of the pre-trained model. Compared with a supervised baseline and the WavLM-based SS model using feature embeddings obtained with the previously released 94K hours trained WavLM, our proposed model obtains 15.9% and 11.2% of relative word error rate (WER) reductions, respectively, for a simulated far-field speech mixture test set. For conversation transcription on real meeting recordings using continuous speech separation, the proposed model achieves 6.8% and 10.6% of relative WER reductions over the purely supervised baseline on AMI and ICSI evaluation sets, respectively, while reducing the computational cost by 38%.