Speech translation: coupling of recognition and translation

By Hermann Ney. 1999

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Xu Tan
2. Tao Qin
3. Ye Jia
4. Yoshua Bengio
5. Dong Yu
6. Li Deng
7. Yu Zhang
8. Jinyu Li

Citations in the corpus

• Deep Learning: Methods and Applications (2014, background): “The final example given here for hybrid deep networks is based on the idea and work of [144, 267], where one task of discrimination (e.”
• Direct speech-to-speech translation with a sequence-to-sequence model (2019, background): “[5, 6] gave MT access to the lattice of the ASR.”
• UWSpeech: Speech to Speech Translation for Unwritten Languages (2020, background): “Previous works on speech translations has evolved from cascaded models [19, 24, 26, 27, 42] to end-to-end models [5, 15, 33, 41, 43], where the text corresponding to speech are leveraged as auxiliary training for better accuracy.”
• A tutorial survey of architectures, algorithms, and applications for deep learning (2014, background): “The final example given here of the hybrid deep architecture is based on the idea and work of [147, 148], where one task of discrimination (speech recognition) produces the output (text) that serves as the input to the second task of discrimination (machine translation).”
• Neural Lattice-to-Sequence Models for Uncertain Inputs (2017, unknown): “(Ney, 1999; Casacuberta et al., 2004).”
• A Scalable Hierarchical Gaussian Process Classifier (2019, background): “For instance, the prediction uncertainty of speech recognition helps improve the performance of a speech translation system [11].”
• MAM: Masked Acoustic Modeling for End-to-End Speech-to-Text Translation (2020, unknown): “…in many scenarios such as international conferences, travels, foreign-language video sub-titling, etc. Conventional cascaded approaches to ST (Ney, 1999; Matusov et al., 2005; Mathias and Byrne, 2006; Berard et al., 2016) first transcribe the speech audio into source language text (ASR) and…”
• On the integration of speech recognition and statistical machine translation (2005, background): “…to minimize the number of sentence errors we maximize the posterior probability of the target language translation given the speech signal (see [2]): Note that we made the natural assumption that the speech signal does not depend on the target sentence and approximated the sum over all…”
• Leveraging Weakly Supervised Data to Improve End-to-end Speech-to-text Translation (2018, methodology): “Early work on speech translation typically used a cascade of an ASR model and an MT model [12, 13, 14], giving the MT model access to the predicted probabilities and uncertainties from the ASR.”
• Speech-Centric Information Processing: An Optimization-Oriented Approach (2013, methodology): “We present an optimization-oriented statistical framework for the overall system design where the interactions between the subsystems in tandem are fully incorporated and where design consistency is established between the optimization objectives and the end-to-end system performance metrics.”
• Toward Robust Neural Machine Translation for Noisy Input Sequences (2017, background): “Finally, prior work has dealt with uncertain inputs from upstream components through explicit representation of the uncertainty, for example by directly translating word lattices produced by the speech recognizer [24, 25, 26, 27].”
• Speech Translation and the End-to-End Promise: Taking Stock of Where We Are (2020, background): “While the desirable solution of full integration over transcripts is intractable (Ney, 1999), approximations are possible.”
• Tied Multitask Learning for Neural Speech Translation (2018, methodology): “For example, Ney (1999) and Matusov et al. (2005) use ASR output lattices as input to translation models, integrating speech recognition uncertainty into the translation model.”
• Sequence-to-Sequence Models Can Directly Translate Foreign Speech (2017, methodology): “Early work on speech translation (ST) [14] – translating audio in one language into text in another – used lattices from an ASR system as inputs to translation models [15, 16], giving the translation model access to the speech recognition uncertainty.”