Skip to search formSkip to main contentSkip to account menu
The Multilingual TEDx Corpus for Speech Recognition and Translation
TLDR
The Multilingual TEDx corpus is a collection of audio recordings from TEDx talks in 8 source languages built to support speech recognition (ASR) and speech translation (ST) research across many non-English source languages.
View PDF on arXiv
Fluent Translations from Disfluent Speech in End-to-End Speech Translation
TLDR
This work uses a sequence-to-sequence model to translate from noisy, disfluent speech to fluent text with disfluencies removed using the recently collected ‘copy-edited’ references for the Fisher Spanish-English dataset.
View on ACL
FINDINGS OF THE IWSLT 2020 EVALUATION CAMPAIGN
TLDR
Each track’s goal, data and evaluation metrics are introduced, and the results of the received submissions are reported.
View on ACL
The IWSLT 2019 Evaluation Campaign
The IWSLT 2019 evaluation campaign featured three tasks: speech translation of (i) TED talks and (ii) How2 instructional videos from English into German and Portuguese, and (iii) text translation of…
View on ACL
Exploring Phoneme-Level Speech Representations for End-to-End Speech Translation
TLDR
This work shows that a naive method to create compressed phoneme-like speech representations is far more effective and efficient for translation than traditional frame-level speech features.
View on ACL
A Language-Independent Approach to Automatic Text Difficulty Assessment for Second-Language Learners
TLDR
A new baseline for language-independent text difficulty assessment applied to the Interagency Language Roundtable (ILR) proficiency scale is introduced and it is demonstrated that reading level assessment is a discriminative problem that is best-suited for regression.
View on ACL
Towards Fluent Translations From Disfluent Speech
TLDR
A corpus of cleaned target data for the Fisher Spanish-English dataset is introduced to compare how different architectures handle disfluency and provide a baseline for removing disfluencies in end-to-end translation.
View on IEEE
Between words and characters: A Brief History of Open-Vocabulary Modeling and Tokenization in NLP
TLDR
This survey connects several lines of work from the pre-neural and neural era, by showing how hybrid approaches of words and characters as well as subwordbased approaches based on learned segmentation have been proposed and evaluated.
View PDF on arXiv
Exploiting Morphological, Grammatical, and Semantic Correlates for Improved Text Difficulty Assessment
TLDR
This work demonstrates that the addition of morphological, information theoretic, and language modeling features to a traditional readability baseline greatly benefits this system's performance.
View on ACL
Relative Positional Encoding for Speech Recognition and Direct Translation
TLDR
This work adapts the relative position encoding scheme to the Speech Transformer, where the key addition is relative distance between input states in the self-attention network, so the network can better adapt to the variable distributions present in speech data.
View PDF on arXiv
...
...
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy, Terms of Service, and Dataset License