Abstract

The language diversity in India's education sector poses a significant challenge, hindering inclusivity. Despite the democratization of knowledge through online educational content, the dominance of English, as the internet's lingua franca, limits accessibility, emphasizing the crucial need for translation into Indian languages. Despite existing Speech-to-Speech Machine Translation (SSMT) technologies, the lack of intonation in these systems gives monotonous translations, leading to a loss of audience interest and disengagement from the content. To address this, our paper introduces a dataset with stress annotations in Indian English and also a Text-to-Speech (TTS) architecture capable of incorporating stress into synthesized speech. This dataset is used for training a stress detection model, which is then used in the SSMT system for detecting stress in the source speech and transferring it into the target language speech. The TTS architecture is based on FastPitch and can modify the variances based on stressed words given. We present an Indian English-to-Hindi SSMT system that can transfer stress and aim to enhance the overall quality and engagement of educational content.

Abstract

India's education sector faces a significant challenge due to its linguistic diversity, hindering inclusivity. The dominance of English on the internet underscores the critical need for translating educational content into Indian languages to enhance accessibility. Although Speech-to-Speech Machine Translation (SSMT) technologies exist, their deficiency in reproducing intonation results in monotonous translations, diminishing audience engagement and interest in the content. To address this issue, this paper demonstrates an SSMT application with a Text-to-Speech (TTS) architecture capable of incorporating stress into synthesized speech to give a more engaging experience. The SSMT pipeline also has components like a stress classifier that captures the stress in the source speech and allows it to be utilized during speech generation. The application takes in a speech file and gives a translated speech file with stress transferred from the source.

Abstract

Identifying keywords in an open-vocabulary context is crucial for personalizing interactions with smart devices. Previous approaches to open vocabulary keyword spotting dependon a shared embedding space created by audio and text encoders. However, these approaches suffer from heterogeneous modality representations (i.e., audio-text mismatch). To address this issue, our proposed framework leverages knowledge acquired from a pre-trained text-to-speech (TTS) system. This knowledge transfer allows for the incorporation of awareness of audio projections into the text representations derived from the text encoder. The performance of the proposed approach is compared with various baseline methods across four different datasets. The robustness of our proposed model is evaluated by assessing its performance across different word lengths and in an Out-of-Vocabulary (OOV) scenario. Additionally, the effectiveness of transfer learning from the TTS system is investigated by analyzing its different intermediate representations. The experimental results indicate that, in the challenging LibriPhrase Hard dataset, the proposed approach outperformed the cross-modality correspondence detector (CMCD) method by a significant improvement of 8.22% in area under the curve (AUC) and 12.56% in equal error rate (EER).

Abstract

Identifying user-defined keywords is crucial for personalizing interactions with smart devices. Previous approaches of user-defined keyword spotting (UDKWS) have relied on short-term spectral features such as mel frequency cepstral coefficients (MFCC) to detect the spoken keyword. However, these features may face challenges in accurately identifying closely related pronunciation of audio-text pairs, due to their limited capability in capturing the temporal dynamics of the speech signal. To address this challenge, we propose to use shifted delta coefficients (SDC) which help in capturing pronunciation variability (transition between connecting phonemes) by incorporating long-term temporal information. The performance of the SDC feature is compared with various baseline features across four different datasets using a cross-attention based end-to-end system. Additionally, various configurations of SDC are explored to find the suitable temporal context for the UDKWS task. The experimental results reveal that the SDC feature outperforms the MFCC baseline feature, exhibiting an improvement of 8.32% in area under the curve (AUC) and 8.69% in terms of equal error rate (EER) on the challenging Libriphrase-hard dataset. Moreover, the proposed approach demonstrated superior performance when compared to state-of-the-art UDKWS techniques.

Abstract

Stuttering is a prevalent speech disorder that affects millions of people worldwide. In this Show and Tell presentation, we demonstrate a novel platform that takes speech samples in English and Kannada to detect and analyze stuttering in patients. The user-friendly interface includes demographic details and speech samples, generating comprehensive reports for different stuttering disfluencies. The platform has four different user types, providing full read-only access for admins and full write access for super admins. Our platform provides valuable assistance for speechlanguage pathologists to evaluate speech samples. The proposed platform supports both live and recorded speech samples and presents a flexible approach to stuttering detection and analysis. Our research demonstrates the potential of technology to improve speech-language pathology for stuttering. Used F-score as a metric for evaluating the models for the stutter detection task.

Abstract

Multilingual Automatic Speech Recognition (ASR) for Indian lan- guages is an obvious technique for leveraging their similarities. We present a detailed analysis of how phonological similarities and differences between languages affect Time Delay Neural Net- work (TDNN) and End-to-End (E2E) ASR. To study this, we select genealogically similar pairs from five Indian languages and train bilingual acoustic models. We compare these against corresponding monolingual acoustic models and find similar phoneme distribu- tions within speech to be the primary factor for improving model performance, with phoneme overlap being secondary. The influ- ence of phonological properties on performance is visible in both cases. Word Error Rate (WER) of E2E decreased by a median of 2.35%, and upto 8.5% when the phonological similarity was great- est. WER of TDNN increased by 11.69% when the similarity was lowest. Thus, it is clear that the choice of supplementary language is important for model performance. Index Terms— Phonetics and Phonology, Analysis, Automatic Speech Recognition, Low-Resource Languages, Multilingual Sys-

Abstract

Hardware accelerators are being designed to offload compute-intensive tasks such as deep neural networks from the CPU to improve the overall performance of an application, specifically on the performance-per-watt metric. Encoder-decoderbased sequence-to-sequence models such as the Transformer model have demonstrated state-of-the-art results in end-to-end automatic speech recognition systems (ASRs). The Transformer model being intensive on memory and computation poses a challenge for an FPGA implementation. This paper proposes an end-to-end architecture to accelerate a Transformer for an ASR system. The host CPU orchestrates the computations from different encoder and decoder stages of the Transformer architecture on the designed hardware accelerator with no necessity for intervening FPGA reconfiguration. The communication latency is hidden by prefetching the weights of the next encoder/decoder block while the current block is being processed. The computation is split across both the Super Logic Regions (SLRs) of the FPGA, mitigating the inter-SLR communication. The proposed design presents an optimal latency, exploiting the available resources. The accelerator design is realized using highlevel synthesis tools and evaluated on an Alveo U-50 FPGA card. The design demonstrates an average speed-up of 32× compared to an Intel Xeon E5-2640 CPU and 8.8× compared to NVIDIA GeForce RTX 3080 Ti Graphics card for a 32-bit floating point single precision model. Index Terms—Automatic speech recognition, Hardware accelerator, Transformer, FPGAs.

Abstract

This work proposes the utilization of a self-supervised pretrained network for developing a Language Identification (LID) system catering to low-resource Indian languages. The framework employed is Wav2vec2.0-XLSR-53, pre-trained on 53k hours of unlabeled speech data. The unsupervised training of the model enables it to learn the acoustic patterns specific to a language. Given that languages share phonetic space, multi-lingual pre-training is instrumental in learning crosslingual information and building systems that cater to low-resource languages. The results showcase a relative improvement of 33.2% over the DNN-A (DNN with attention) model and 19.04% over Dense Resnets for the Language Identification task on the IIITH-ILSC database using the proposed features.

Abstract

Due to the lack of a large annotated speech corpus, many low-resource Indian languages struggle to utilize recent advancements in deep neural network architectures for Automatic Speech Recognition (ASR) tasks. Collecting large-scale databases is an expensive and time-consuming task. Current approaches lack extensive traditional expert-based data acquisition guidelines, as they are tedious and complex. In this work, we present the International Institute of Information Technology Hyderabad-Crowd Sourced Telugu Database (IIITH-CSTD), a Telugu corpus collected through crowdsourcing

Abstract

Voice activity detection (VAD), is a signal processing technique used to determine whether a given speech signal contains voiced or unvoiced segments. VAD is used in various applications such as Speech Coding, Voice Controlled Systems, speech feature extraction, etc. For example, in Adaptive multirate (AMR) speech coding, VAD is used as an efficient way of coding different speech frames at different bit rates. In this paper, we implemented the application of a Zero-Phase Zero Frequency Resonator (ZP-ZFR) as VAD on hardware. ZP-ZFR is an Infinite Impulse Response (IIR) filter that offers the advantage of requiring a lower filter order, making it suitable for hardware implementation. The proposed system is implemented on the TIMIT database using the Nexys Video Artix-7 FPGA board. The hardware design is carried out using Vivado 2021.1, a popular tool for FPGA development. The Hardware Description Language (HDL) used for implementation is Verilog. The proposed system achieves good performance with low complexity. Therefore this work is implemented on hardware, which can be used in various applications. Index Terms—Voice Activity Detection (VAD), Zero-Phase Zero Frequency Resonator (ZP-ZFR), Zero Frequency Filter (ZFF), MATLAB, Field Programmable Gate Array (FPGA).