Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userJointly Trained Transformers models for Spoken Language Translation
72
0
0.0
(
0
)
Added by
Hari Krishna Vydana Mr
Publication date
2020
fields
Informatics Engineering
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
Conventional spoken language translation (SLT) systems are pipeline based systems, where we have an Automatic Speech Recognition (ASR) system to convert the modality of source from speech to text and a Machine Translation (MT) systems to translate source text to text in target language. Recent progress in the sequence-sequence architectures have reduced the performance gap between the pipeline based SLT systems (cascaded ASR-MT) and End-to-End approaches. Though End-to-End and cascaded ASR-MT systems are reaching to the comparable levels of performances, we can see a large performance gap using the ASR hypothesis and oracle text w.r.t MT models. This performance gap indicates that the MT systems are prone to large performance degradation due to noisy ASR hypothesis as opposed to oracle text transcript. In this work this degradation in the performance is reduced by creating an end to-end differentiable pipeline between the ASR and MT systems. In this work, we train SLT systems with ASR objective as an auxiliary loss and both the networks are connected through the neural hidden representations. This train ing would have an End-to-End differentiable path w.r.t to the final objective function as well as utilize the ASR objective for better performance of the SLT systems. This architecture has improved from BLEU from 36.8 to 44.5. Due to the Multi-task training the model also generates the ASR hypothesis which are used by a pre-trained MT model. Combining the proposed systems with the MT model has increased the BLEU score by 1. All the experiments are reported on English-Portuguese speech translation task using How2 corpus. The final BLEU score is on-par with the best speech translation system on How2 dataset with no additional training data and language model and much less parameters.
rate research
Read More
We propose spoken sentence embeddings which capture both acoustic and linguistic content. While existing works operate at the character, phoneme, or word level, our method learns long-term dependencies by modeling speech at the sentence level. Formulated as an audio-linguistic multitask learning problem, our encoder-decoder model simultaneously reconstructs acoustic and natural language features from audio. Our results show that spoken sentence embeddings outperform phoneme and word-level baselines on speech recognition and emotion recognition tasks. Ablation studies show that our embeddings can better model high-level acoustic concepts while retaining linguistic content. Overall, our work illustrates the viability of generic, multi-modal sentence embeddings for spoken language understanding.
In this paper, we address the task of spoken language understanding. We present a method for translating spoken sentences from one language into spoken sentences in another language. Given spectrogram-spectrogram pairs, our model can be trained completely from scratch to translate unseen sentences. Our method consists of a pyramidal-bidirectional recurrent network combined with a convolutional network to output sentence-level spectrograms in the target language. Empirically, our model achieves competitive performance with state-of-the-art methods on multiple languages and can generalize to unseen speakers.
In this paper, we present a streaming end-to-end speech recognition model based on Monotonic Chunkwise Attention (MoCha) jointly trained with enhancement layers. Even though the MoCha attention enables streaming speech recognition with recognition accuracy comparable to a full attention-based approach, training this model is sensitive to various factors such as the difficulty of training examples, hyper-parameters, and so on. Because of these issues, speech recognition accuracy of a MoCha-based model for clean speech drops significantly when a multi-style training approach is applied. Inspired by Curriculum Learning [1], we introduce two training strategies: Gradual Application of Enhanced Features (GAEF) and Gradual Reduction of Enhanced Loss (GREL). With GAEF, the model is initially trained using clean features. Subsequently, the portion of outputs from the enhancement layers gradually increases. With GREL, the portion of the Mean Squared Error (MSE) loss for the enhanced output gradually reduces as training proceeds. In experimental results on the LibriSpeech corpus and noisy far-field test sets, the proposed model with GAEF-GREL training strategies shows significantly better results than the conventional multi-style training approach.
Spoken language translation (SLT) is becoming more important in the increasingly globalized world, both from a social and economic point of view. It is one of the major challenges for automatic speech recognition (ASR) and machine translation (MT), driving intense research activities in these areas. While past research in SLT, due to technology limitations, dealt mostly with speech recorded under controlled conditions, todays major challenge is the translation of spoken language as it can be found in real life. Considered application scenarios range from portable translators for tourists, lectures and presentations translation, to broadcast news and shows with live captioning. We would like to present PJIITs experiences in the SLT gained from the Eu-Bridge 7th framework project and the U-Star consortium activities for the Polish/English language pair. Presented research concentrates on ASR adaptation for Polish (state-of-the-art acoustic models: DBN-BLSTM training, Kaldi: LDA+MLLT+SAT+MMI), language modeling for ASR & MT (text normalization, RNN-based LMs, n-gram model domain interpolation) and statistical translation techniques (hierarchical models, factored translation models, automatic casing and punctuation, comparable and bilingual corpora preparation). While results for the well-defined domains (phrases for travelers, parliament speeches, medical documentation, movie subtitling) are very encouraging, less defined domains (presentation, lectures) still form a challenge. Our progress in the IWSLT TED task (MT only) will be presented, as well as current progress in the Polish ASR.
In this paper we describe how the translation methodology adopted for the Spoken Language Translator (SLT) addresses the characteristics of the speech translation task in a context where it is essential to achieve easy customization to new languages and new domains. We then discuss the issues that arise in any attempt to evaluate a speech translator, and present the results of such an evaluation carried out on SLT for several language pairs.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
