Machine learning (ML) is widely used to moderate online content. Despite its scalability relative to human moderation, the use of ML introduces unique challenges to content moderation. One such challenge is predictive multiplicity: multiple competing models for content classification may perform equally well on average, yet assign conflicting predictions to the same content. This multiplicity can result from seemingly innocuous choices during model development, such as random seed selection for parameter initialization. We experimentally demonstrate how content moderation tools can arbitrarily classify samples as toxic, leading to arbitrary restrictions on speech. We discuss these findings in terms of human rights set out by the International Covenant on Civil and Political Rights (ICCPR), namely freedom of expression, non-discrimination, and procedural justice. We analyze (i) the extent of predictive multiplicity among state-of-the-art LLMs used for detecting toxic content; (ii) the disparate impact of this arbitrariness across social groups; and (iii) how model multiplicity compares to unambiguous human classifications. Our findings indicate that the up-scaled algorithmic moderation risks legitimizing an algorithmic leviathan, where an algorithm disproportionately manages human rights. To mitigate such risks, our study underscores the need to identify and increase the transparency of arbitrariness in content moderation applications. Since algorithmic content moderation is being fueled by pressing social concerns, such as disinformation and hate speech, our discussion on harms raises concerns relevant to policy debates. Our findings also contribute to content moderation and intermediary liability laws being discussed and passed in many countries, such as the Digital Services Act in the European Union, the Online Safety Act in the United Kingdom, and the Fake News Bill in Brazil.
In this paper, we examine the research conducted in the field of Nepali Automatic Speech Recognition (ASR). The primary objective of this survey is to conduct a comprehensive review of the works on Nepali Automatic Speech Recognition Systems completed to date, explore the different datasets used, examine the technology utilized, and take account of the obstacles encountered in implementing the Nepali ASR system. In tandem with the global trends of ever-increasing research on speech recognition based research, the number of Nepalese ASR-related projects are also growing. Nevertheless, the investigation of language and acoustic models of the Nepali language has not received adequate attention compared to languages that possess ample resources. In this context, we provide a framework as well as directions for future investigations.
Neural Text-to-speech (TTS) synthesis is a powerful technology that can generate speech using neural networks. One of the most remarkable features of TTS synthesis is its capability to produce speech in the voice of different speakers. This paper introduces voice cloning and speech synthesis https://pypi.org/project/voice-cloning/ an open-source python package for helping speech disorders to communicate more effectively as well as for professionals seeking to integrate voice cloning or speech synthesis capabilities into their projects. This package aims to generate synthetic speech that sounds like the natural voice of an individual, but it does not replace the natural human voice. The architecture of the system comprises a speaker verification system, a synthesizer, a vocoder, and noise reduction. Speaker verification system trained on a varied set of speakers to achieve optimal generalization performance without relying on transcriptions. Synthesizer is trained using both audio and transcriptions that generate Mel spectrogram from a text and vocoder which converts the generated Mel Spectrogram into corresponding audio signal. Then the audio signal is processed by a noise reduction algorithm to eliminate unwanted noise and enhance speech clarity. The performance of synthesized speech from seen and unseen speakers are then evaluated using subjective and objective evaluation such as Mean Opinion Score (MOS), Gross Pitch Error (GPE), and Spectral distortion (SD). The model can create speech in distinct voices by including speaker characteristics that are chosen randomly.
In spoken dialogue, even if two current turns are the same sentence, their responses might still differ when they are spoken in different styles. The spoken styles, containing paralinguistic and prosodic information, mark the most significant difference between text and speech modality. When using text-only LLMs to model spoken dialogue, text-only LLMs cannot give different responses based on the speaking style of the current turn. In this paper, we focus on enabling LLMs to listen to the speaking styles and respond properly. Our goal is to teach the LLM that "even if the sentences are identical if they are spoken in different styles, their corresponding responses might be different". Since there is no suitable dataset for achieving this goal, we collect a speech-to-speech dataset, StyleTalk, with the following desired characteristics: when two current speeches have the same content but are spoken in different styles, their responses will be different. To teach LLMs to understand and respond properly to the speaking styles, we propose the Spoken-LLM framework that can model the linguistic content and the speaking styles. We train Spoken-LLM using the StyleTalk dataset and devise a two-stage training pipeline to help the Spoken-LLM better learn the speaking styles. Based on extensive experiments, we show that Spoken-LLM outperforms text-only baselines and prior speech LLMs methods.
Despite the predominance of English in their training data, English-centric Large Language Models (LLMs) like GPT-3 and LLaMA display a remarkable ability to perform multilingual tasks, raising questions about the depth and nature of their cross-lingual capabilities. This paper introduces the decomposed prompting approach to probe the linguistic structure understanding of these LLMs in sequence labeling tasks. Diverging from the single text-to-text prompt, our method generates for each token of the input sentence an individual prompt which asks for its linguistic label. We assess our method on the Universal Dependencies part-of-speech tagging dataset for 38 languages, utilizing both English-centric and multilingual LLMs. Our findings show that decomposed prompting surpasses the iterative prompting baseline in efficacy and efficiency under zero- and few-shot settings. Further analysis reveals the influence of evaluation methods and the use of instructions in prompts. Our multilingual investigation shows that English-centric language models perform better on average than multilingual models. Our study offers insights into the multilingual transferability of English-centric LLMs, contributing to the understanding of their multilingual linguistic knowledge.
Speech intelligibility can be affected by multiple factors, such as noisy environments, channel distortions or physiological issues. In this work, we deal with the problem of automatic prediction of the speech intelligibility level in this latter case. Starting from our previous work, a non-intrusive system based on LSTM networks with attention mechanism designed for this task, we present two main contributions. In the first one, it is proposed the use of per-frame modulation spectrograms as input features, instead of compact representations derived from them that discard important temporal information. In the second one, two different strategies for the combination of per-frame acoustic log-mel and modulation spectrograms into the LSTM framework are explored: at decision level or late fusion and at utterance level or Weighted-Pooling (WP) fusion. The proposed models are evaluated with the UA-Speech database that contains dysarthric speech with different degrees of severity. On the one hand, results show that attentional LSTM networks are able to adequately modeling the modulation spectrograms sequences producing similar classification rates as in the case of log-mel spectrograms. On the other hand, both combination strategies, late and WP fusion, outperform the single-feature systems, suggesting that per-frame log-mel and modulation spectrograms carry complementary information for the task of speech intelligibility prediction, than can be effectively exploited by the LSTM-based architectures, being the system with the WP fusion strategy and Attention-Pooling the one that achieves best results.
Text-to-speech models trained on large-scale datasets have demonstrated impressive in-context learning capabilities and naturalness. However, control of speaker identity and style in these models typically requires conditioning on reference speech recordings, limiting creative applications. Alternatively, natural language prompting of speaker identity and style has demonstrated promising results and provides an intuitive method of control. However, reliance on human-labeled descriptions prevents scaling to large datasets. Our work bridges the gap between these two approaches. We propose a scalable method for labeling various aspects of speaker identity, style, and recording conditions. We then apply this method to a 45k hour dataset, which we use to train a speech language model. Furthermore, we propose simple methods for increasing audio fidelity, significantly outperforming recent work despite relying entirely on found data. Our results demonstrate high-fidelity speech generation in a diverse range of accents, prosodic styles, channel conditions, and acoustic conditions, all accomplished with a single model and intuitive natural language conditioning. Audio samples can be heard at https://text-description-to-speech.com/.
We present MunTTS, an end-to-end text-to-speech (TTS) system specifically for Mundari, a low-resource Indian language of the Austo-Asiatic family. Our work addresses the gap in linguistic technology for underrepresented languages by collecting and processing data to build a speech synthesis system. We begin our study by gathering a substantial dataset of Mundari text and speech and train end-to-end speech models. We also delve into the methods used for training our models, ensuring they are efficient and effective despite the data constraints. We evaluate our system with native speakers and objective metrics, demonstrating its potential as a tool for preserving and promoting the Mundari language in the digital age.
Relating speech to EEG holds considerable importance but is challenging. In this study, a deep convolutional network was employed to extract spatiotemporal features from EEG data. Self-supervised speech representation and contextual text embedding were used as speech features. Contrastive learning was used to relate EEG features to speech features. The experimental results demonstrate the benefits of using self-supervised speech representation and contextual text embedding. Through feature fusion and model ensemble, an accuracy of 60.29% was achieved, and the performance was ranked as No.2 in Task 1 of the Auditory EEG Challenge (ICASSP 2024). The code to implement our work is available on Github: https://github.com/bobwangPKU/EEG-Stimulus-Match-Mismatch.
Supervised models for speech enhancement are trained using artificially generated mixtures of clean speech and noise signals. However, the synthetic training conditions may not accurately reflect real-world conditions encountered during testing. This discrepancy can result in poor performance when the test domain significantly differs from the synthetic training domain. To tackle this issue, the UDASE task of the 7th CHiME challenge aimed to leverage real-world noisy speech recordings from the test domain for unsupervised domain adaptation of speech enhancement models. Specifically, this test domain corresponds to the CHiME-5 dataset, characterized by real multi-speaker and conversational speech recordings made in noisy and reverberant domestic environments, for which ground-truth clean speech signals are not available. In this paper, we present the objective and subjective evaluations of the systems that were submitted to the CHiME-7 UDASE task, and we provide an analysis of the results. This analysis reveals a limited correlation between subjective ratings and several supervised nonintrusive performance metrics recently proposed for speech enhancement. Conversely, the results suggest that more traditional intrusive objective metrics can be used for in-domain performance evaluation using the reverberant LibriCHiME-5 dataset developed for the challenge. The subjective evaluation indicates that all systems successfully reduced the background noise, but always at the expense of increased distortion. Out of the four speech enhancement methods evaluated subjectively, only one demonstrated an improvement in overall quality compared to the unprocessed noisy speech, highlighting the difficulty of the task. The tools and audio material created for the CHiME-7 UDASE task are shared with the community.