Recent advancements in joint speech-text pre-training have significantly advanced the processing of natural language. However, a key limitation is their reliance on parallel speech-text data, posing challenges due to data accessibility. Addressing this, our paper introduces an innovative framework for jointly performing speech and text processing without parallel corpora during pre-training but only downstream. Utilizing pre-trained unimodal models, we extract distinct representations for speech and text, aligning them effectively in a newly defined space using a multi-level contrastive learning mechanism. A unique swap reconstruction mechanism enhances the alignment and is followed by fusion via a multi-head mechanism, seamlessly merging modality-invariant and modality-specific representations. Testing for emotion recognition (SLU task) and idiom usage detection (NLU task) demonstrates robust performance, with commendable robustness to noise in text or speech data.
论文检索
输入标题、作者或关键词，从 2,480 篇学术成果中精准定位
Pre-trained speech models, such as wav2vec 2.0, have significantly advanced speech-related tasks, including speech recognition and translation. However, their applicability in streaming scenarios is limited because these models are trained on complete utterances, leading to a mismatch with incremental streaming inputs. This paper identifies three critical design aspects within the architecture of wav2vec 2.0 and proposes a novel model, wav2vec-S, which incorporates simple modifications to ensure consistent speech representations during both training and inference phases for streaming speech inputs. Furthermore, we demonstrate that wav2vec-S models can be efficiently adapted from pre-trained wav2vec 2.0 models through continued pre-training and effectively finetuned to meet various latency requirements in downstream applications. Experiments on speech recognition and translation tasks show that wav2vec-S outperforms strong baseline models and achieves a superior balance between quality and latency.
We present INDICVOICES, a dataset of natural and spontaneous speech containing a total of 7348 hours of read (9%), extempore (74%) and conversational (17%) audio from 16237 speakers covering 145 Indian districts and 22 languages. Of these 7348 hours, 1639 hours have already been transcribed, with a median of 73 hours per language. Through this paper, we share our journey of capturing the cultural, linguistic and demographic diversity of India to create a one-of-its-kind inclusive and representative dataset. More specifically, we share an open-source blueprint for data collection at scale comprising of standardised protocols, centralised tools, a repository of engaging questions, prompts and conversation scenarios spanning multiple domains and topics of interest, quality control mechanisms, comprehensive transcription guidelines and transcription tools. We hope that this open source blueprint will serve as a comprehensive starter kit for data collection efforts in other multilingual regions of the world. Using INDICVOICES, we build IndicASR, the first ASR model to support all the 22 languages listed in the 8th schedule of the Constitution of India.
The ever-growing presence of hate speech on social network services and other online platforms not only fuels online harassment but also presents a growing challenge for hate speech detection. As this task is akin to binary classification, one of the promising approaches for hate speech detection is the utilization of contrastive learning. Recent studies suggest that classifying hateful posts in just a binary manner may not adequately address the nuanced task of detecting implicit hate speech. This challenge is largely due to the subtle nature and context dependency of such pejorative remarks. Previous studies proposed a modified contrastive learning approach equipped with additional aids such as human-written implications or machine-generated augmented data for better implicit hate speech detection. While this approach can potentially enhance the overall performance by its additional data in general, it runs the risk of overfitting as well as heightened cost and time to obtain. These drawbacks serve as motivation for us to design a methodology that is not dependent on human-written or machine-generated augmented data for training. We propose a straightforward, yet effective, clustering-based contrastive learning approach that leverages the shared semantics among the data.
Emotion plays a crucial role in human conversation. This paper underscores the significance of considering emotion in speech translation. We present the MELD-ST dataset for the emotion-aware speech translation task, comprising English-to-Japanese and English-to-German language pairs. Each language pair includes about 10,000 utterances annotated with emotion labels from the MELD dataset. Baseline experiments using the SeamlessM4T model on the dataset indicate that fine-tuning with emotion labels can enhance translation performance in some settings, highlighting the need for further research in emotion-aware speech translation systems.
Integrated Speech and Large Language Models (SLMs) that can follow speech instructions and generate relevant text responses have gained popularity lately. However, the safety and robustness of these models remains largely unclear. In this work, we investigate the potential vulnerabilities of such instruction-following speech-language models to adversarial attacks and jailbreaking. Specifically, we design algorithms that can generate adversarial examples to jailbreak SLMs in both white-box and black-box attack settings without human involvement. Additionally, we propose countermeasures to thwart such jailbreaking attacks. Our models, trained on dialog data with speech instructions, achieve state-of-the-art performance on spoken question-answering task, scoring over 80% on both safety and helpfulness metrics. Despite safety guardrails, experiments on jailbreaking demonstrate the vulnerability of SLMs to adversarial perturbations and transfer attacks, with average attack success rates of 90% and 10% respectively when evaluated on a dataset of carefully designed harmful questions spanning 12 different toxic categories. However, we demonstrate that our proposed countermeasures reduce the attack success significantly.
Direct speech-to-speech translation achieves high-quality results through the introduction of discrete units obtained from self-supervised learning. However, talking head translation, converting audio-visual speech (i.e., talking head video) from one language into another, still confronts several challenges compared to audio speech: (1) Existing methods invariably rely on cascading, synthesizing via both audio and text, resulting in delays and cascading errors. (2) Talking head translation has a limited set of reference frames. If the generated translation exceeds the length of the original speech, the video sequence needs to be supplemented by repeating frames, leading to jarring video transitions. In this work, we propose a model for talking head translation, TransFace, which can directly translate audio-visual speech into audio-visual speech in other languages. It consists of a speech-to-unit translation model to convert audio speech into discrete units and a unit-based audio-visual speech synthesizer, Unit2Lip, to re-synthesize synchronized audio-visual speech from discrete units in parallel. Furthermore, we introduce a Bounded Duration Predictor, ensuring isometric talking head translation and preventing duplicate reference frames. Experiments demonstrate that Unit2Lip significantly improves synchronization and boosts inference speed by a factor of 4.35 on LRS2. Additionally, TransFace achieves impressive BLEU scores of 61.93 and 47.55 for Es-En and Fr-En on LRS3-T and 100% isochronous translations. The samples are available at https://transface-demo.github.io .
Recent advances in large language models (LLMs) have led to significant success in using LLMs as agents. Nevertheless, a common assumption that LLMs always process honest information neglects the widespread deceptive or misleading content in human and AI-generated material. This oversight might expose LLMs to malicious manipulations. To enhance LLMs’ ability to identify and counteract deceptive information, in this paper, inspired by humans’ recursive thinking and perspective-taking, we introduce a novel cognitive framework, Recursive Contemplation (ReCon). ReCon combines formulation and refinement contemplation processes; formulation contemplation produces initial thoughts and speech, while refinement contemplation further polishes them. Additionally, we incorporate first-order and second-order perspective transitions into these processes respectively. Specifically, the first-order allows an LLM agent to infer others’ mental states, and the second-order involves understanding how others perceive the agent’s mental state. After integrating ReCon with various LLMs, extensive experiment results from the Avalon game and BigTom benchmark indicate ReCon’s efficacy in aiding LLMs to discern and maneuver around deceptive information without extra fine-tuning and data. Finally, we demonstrate ReCon’s scaling trend with model parameters, and explore the current limitations of LLMs in terms of safety and reasoning, potentially furnishing insights for subsequent research. Our project page can be found at https://shenzhi-wang.github.io/avalon_recon.
Speech-to-singing voice conversion (STS) task always suffers from data scarcity, because it requires paired speech and singing data. Compounding this issue are the challenges of content-pitch alignment and the suboptimal quality of generated outputs, presenting significant hurdles in STS research. This paper presents SVPT, an STS approach boosted by a self-supervised singing voice pre-training model.We leverage spoken language model techniques to tackle the rhythm alignment problem and the in-context learning capability to achieve zero-shot conversion. We adopt discrete-unit random resampling and pitch corruption strategies, enabling training with unpaired singing data and thus mitigating the issue of data scarcity. SVPT also serves as an effective backbone for singing voice synthesis (SVS), offering insights into scaling up SVS models. Experimental results indicate that SVPT delivers notable improvements in both STS and SVS endeavors. Audio samples are available at https://speech2sing.github.io.
Speech-to-text (S2T) generation systems frequently face challenges in low-resource scenarios, primarily due to the lack of extensive labeled datasets. One emerging solution is constructing virtual training samples by interpolating inputs and labels, which has notably enhanced system generalization in other domains. Despite its potential, this technique’s application in S2T tasks has remained under-explored. In this paper, we delve into the utility of interpolation augmentation, guided by several pivotal questions. Our findings reveal that employing an appropriate strategy in interpolation augmentation significantly enhances performance across diverse tasks, architectures, and data scales, offering a promising avenue for more robust S2T systems in resource-constrained settings.
Dialects introduce syntactic and lexical variations in language that occur in regional or social groups. Most NLP methods are not sensitive to such variations. This may lead to unfair behavior of the methods, conveying negative bias towards dialect speakers. While previous work has studied dialect-related fairness for aspects like hate speech, other aspects of biased language, such as lewdness, remain fully unexplored. To fill this gap, we investigate performance disparities between dialects in the detection of five aspects of biased language and how to mitigate them. To alleviate bias, we present a multitask learning approach that models dialect language as an auxiliary task to incorporate syntactic and lexical variations. In our experiments with African-American English dialect, we provide empirical evidence that complementing common learning approaches with dialect modeling improves their fairness. Furthermore, the results suggest that multitask learning achieves state-of-the-art performance and helps to detect properties of biased language more reliably.
Direct speech-to-speech translation (S2ST) has achieved impressive translation quality, but it often faces the challenge of slow decoding due to the considerable length of speech sequences. Recently, some research has turned to non-autoregressive (NAR) models to expedite decoding, yet the translation quality typically lags behind autoregressive (AR) models significantly. In this paper, we investigate the performance of CTC-based NAR models in S2ST, as these models have shown impressive results in machine translation. Experimental results demonstrate that by combining pretraining, knowledge distillation, and advanced NAR training techniques such as glancing training and non-monotonic latent alignments, CTC-based NAR models achieve translation quality comparable to the AR model, while preserving up to 26.81\times decoding speedup.
Semantic Role Labeling (SRL), crucial for understanding semantic relationships in sentences, has traditionally focused on text-based input. However, the increasing use of voice assistants and the need for hands-free interaction have highlighted the importance of SRL from speech.SRL from speech can be accomplished via a two-step pipeline directly: transcribing speech to text via Automatic Speech Recognition (ASR) and then applying text-based SRL, which could lead to error propagation and loss of useful acoustic features.Addressing these challenges, we present the first end-to-end approach for SRL from speech, integrating ASR and SRL in a joint-learning framework, focusing on the Chinese language. By employing a Stright-Through Gumbel-Softmax module for connecting ASR and SRL models, it enables gradient back-propagation and joint optimization, enhancing robustness and effectiveness.Experiments on the Chinese Proposition Bank 1.0 (CPB1.0) and a newly annotated dataset AS-SRL based on AISHELL-1 demonstrate the superiority of the end-to-end model over traditional pipelines, with significantly improved performance.
The translation of gender-neutral person-referring terms (e.g.,the students) is often non-trivial.Translating from English into German poses an interesting case—in German, person-referring nouns are usually gender-specific, and if the gender of the referent(s) is unknown or diverse, the generic masculine (die Studenten (m.)) is commonly used. This solution, however, reduces the visibility of other genders, such as women and non-binary people. To counteract gender discrimination, a societal movement towards using gender-fair language exists (e.g., by adopting neosystems). However, gender-fair German is currently barely supported in machine translation (MT), requiring post-editing or manual translations. We address this research gap by studying gender-fair language in English-to-German MT. Concretely, we enrich a community-created gender-fair language dictionary and sample multi-sentence test instances from encyclopedic text and parliamentary speeches.Using these novel resources, we conduct the first benchmark study involving two commercial systems and six neural MT models for translating words in isolation and natural contexts across two domains. Our findings show that most systems produce mainly masculine forms and rarely gender-neutral variants, highlighting the need for future research. We release code and data at https://github.com/g8a9/building-bridges-gender-fair-german-mt.
Sign language understanding has made significant strides; however, there is still no viable solution for generating sign sequences directlyfrom entire spoken content, e.g., text or speech. In this paper, we propose a unified framework for continuous sign language production, easing communication between sign and non-sign language users. In particular, a sequence diffusion model, utilizing embeddings extracted from text or speech, is crafted to generate sign predictions step by step. Moreover, by creating a joint embedding space for text, audio, and sign, we bind these modalities and leverage the semantic consistency among them to provide informative feedback for the model training. This embedding-consistency learning strategy minimizes the reliance on sign triplets and ensures continuous model refinement, evenwith a missing audio modality. Experiments on How2Sign and PHOENIX14T datasets demonstrate that our model achieves competitive performance in sign language production.
We introduces ***LLaST***, a framework for building high-performance Large Language model based Speech-to-text Translation systems. We address the limitations of end-to-end speech translation (E2E ST) models by exploring model architecture design and optimization techniques tailored for LLMs. Our approach includes LLM-based speech translation architecture design, ASR-augmented training, multilingual data augmentation, and dual-LoRA optimization. Our approach demonstrates superior performance on the CoVoST-2 benchmark and showcases exceptional scaling capabilities powered by LLMs.We believe this effective method will serve as a strong baseline for speech translation and provide insights for futureimprovements of the LLM-based speech translation framework.
Given a script, the challenge in Movie Dubbing (Visual Voice Cloning, V2C) is to generate speech that aligns well with the video in both time and emotion, based on the tone of a reference audio track. Existing state-of-the-art V2C models break the phonemes in the script according to the divisions between video frames, which solves the temporal alignment problem but leads to incomplete phoneme pronunciation and poor identity stability. To address this problem, we propose StyleDubber, which switches dubbing learning from the frame level to phoneme level. It contains three main components: (1) A multimodal style adaptor operating at the phoneme level to learn pronunciation style from the reference audio, and generate intermediate representations informed by the facial emotion presented in the video; (2) An utterance-level style learning module, which guides both the mel-spectrogram decoding and the refining processes from the intermediate embeddings to improve the overall style expression; And (3) a phoneme-guided lip aligner to maintain lip sync. Extensive experiments on two of the primary benchmarks, V2C and Grid, demonstrate the favorable performance of the proposed method as compared to the current state-of-the-art. The code will be made available at https://github.com/GalaxyCong/StyleDubber.
Language models (LMs) are known to represent the perspectives of some social groups better than others, which may impact their performance, especially on subjective tasks such as content moderation and hate speech detection. To explore how LMs represent different perspectives, existing research focused on positional alignment, i.e., how closely the models mimic the opinions and stances of different groups, e.g., liberals or conservatives. However, human communication also encompasses emotional and moral dimensions. We define the problem of affective alignment, which measures how LMs’ emotional and moral tone represents those of different groups. By comparing the affect of responses generated by 36 LMs to the affect of Twitter messages written by two ideological groups, we observe significant misalignment of LMs with both ideological groups. This misalignment is larger than the partisan divide in the U.S. Even after steering the LMs towards specific ideological perspectives, the misalignment and liberal tendencies of the model persist, suggesting a systemic bias within LMs.
Recently, advancements in large language models (LLMs) have shown an unprecedented ability across various language tasks. This paper investigates the potential application of LLMs to slot filling with noisy ASR transcriptions, via both in-context learning and task-specific fine-tuning. Dedicated prompt designs and noise-robust LoRA fine-tuning are proposed to improve the robustness of LLMs for slot filling with noisy ASR transcriptions. Moreover, a linearised knowledge injection (LKI) scheme is also proposed to integrate dynamic external knowledge into LLMs. Experiments were performed on SLURP to quantify the performance of LLMs, including GPT-3.5-turbo, GPT-4, LLaMA-13B, LLaMA-2-13B and Vicuna-13B (v1.1 and v1.5) with different ASR error rates. The use of the noise-robust fine-tuning together with LKI for Vicuna-13B-v1.5 achieved 6.7% and 17.6% absolute SLU-F1 improvements compared to a fully fine-tuned Flan-T5-XL model on the limited data setup and the zero-shot setup respectively.
Social media platforms are hubs for multimodal information exchange, encompassing text, images, and videos, making it challenging for machines to comprehend the information or emotions associated with interactions in online spaces. Multimodal Large Language Models (MLLMs) have emerged as a promising solution to address these challenges, yet struggle with accurately interpreting human emotions and complex contents like misinformation. This paper introduces MM-Soc, a comprehensive benchmark designed to evaluate MLLMs’ understanding of multimodal social media content. MM-Soc compiles prominent multimodal datasets and incorporates a novel large-scale YouTube tagging dataset, targeting a range of tasks from misinformation detection, hate speech detection, and social context generation. Through our exhaustive evaluation on ten size-variants of four open-source MLLMs, we have identified significant performance disparities, highlighting the need for advancements in models’ social understanding capabilities. Our analysis reveals that, in a zero-shot setting, various types of MLLMs generally exhibit difficulties in handling social media tasks. However, MLLMs demonstrate performance improvements post fine-tuning, suggesting potential pathways for improvement.