WIT3: Web Inventory of Transcribed and Translated Talks

By Mauro Cettolo, Christian Girardi, Marcello Federico. 2012

732 citations recorded by the source corpus.

Subfields: Natural Language Processing, Evaluation & Benchmarks

How researchers see this paper

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

In Natural Language Processing

In Evaluation & Benchmarks

Citations in the corpus

Multilingual Denoising Pre-training for Neural Machine Translation (2020, methodology): “The Zh-En dataset is from IWSLT 2014 and 2015 (Cettolo et al., 2012, 2015).”
Improving Neural Machine Translation Models with Monolingual Data (2015, methodology): “As indomain training data, IWSLT provides the WIT3 parallel corpus (Cettolo et al., 2012), which also consists of TED talks.”
Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges (2019, methodology): “The most commonly used datasets include: ( i ) TED talks (Cettolo et al., 2012) which includes 59 languages (Qi et al., 2018) with around 3k to 200k sentence pairs per Figure 1: Per language pair data distribution of the training dataset used for our multilingual experiments.”
Deep double descent: where bigger models and more data hurt (2019, positive): “The IWSLT ’14 German to English dataset contains TED Talks as described in Cettolo et al. (2012).”
What do Neural Machine Translation Models Learn about Morphology? (2017, methodology): “Our translation models are trained on the WIT3 corpus of TED talks (Cettolo et al., 2012; Cettolo, 2016) made available for IWSLT 2016.”
On Using Monolingual Corpora in Neural Machine Translation (2015, methodology): “We used the WIT parallel corpus (Cettolo et al., 2012) and SETimes parallel corpus made available as a part of IWSLT’14 (machine translation track).”
Synthetic and Natural Noise Both Break Neural Machine Translation (2017, background): “Since we do not have access to a parallel corpus with natural noise, we instead harvest naturally occurring errors (typos, misspellings, etc.) from available corpora of edits to build a look-up table of possible lexical replacements.”
Neural Semantic Encoders (2016, methodology): “er NSEs access a shared memory. The memory is encoded by the ﬁrst NSEs and then read/written by the decoder NSEs. We used the English-German translation corpus from the IWSLT 2014 evaluation campaign [35]. The corpus consists of sentence-aligned translation of TED talks. The data was pre-processed a”
Toward Multilingual Neural Machine Translation with Universal Encoder and Decoder (2016, methodology): “We choose WIT3’s4 TED corpus[13] as the basis of our experiments since it might be the only highquality parallel data of many low-resourced language pairs.”
Regularized Training Objective for Continued Training for Domain Adaptation in Neural Machine Translation (2018, neutral): “9 which was extracted from WIT 3 (Cettolo et al., 2012).”
Compression of Neural Machine Translation Models via Pruning (2016, methodology): “To test the generalizability of our results, we also test our pruning approach on a smaller, nonstate-of-the-art NMT model trained on the WIT3 Vietnamese-English dataset (Cettolo et al., 2012), which consists of 133,000 sentence pairs.”
Multilingual Models for Compositional Distributed Semantics (2014, methodology): “Second, we developed a massively multilingual corpus based on the TED corpus 2 for IWSLT 2013 (Cettolo et al., 2012).”
Controllable Invariance through Adversarial Feature Learning (2017, methodology): “to 1 in our experiments. Multi-lingual Machine Translation For the multi-lingual machine translation task we use French to English (fr-en) and German to English (de-en) pairs from IWSLT 2015 dataset [Cettolo et al., 2012]. There are 198;435pairs of fr-en sentences and 188;661pairs of de-en sentences”
Dynamic Data Selection for Neural Machine Translation (2017, background): “, 2016) constructed from OpenSubtitles (Lison and Tiedemann, 2016), (iii) TED talks (Cettolo et al., 2012), and (iv) WMT news.”
Regularization techniques for fine-tuning in neural machine translation (2017, methodology): “We evaluate transfer learning on test sets from the IWSLT shared translation task (Cettolo et al., 2012).”
Learning Joint Multilingual Sentence Representations with Neural Machine Translation (2017, methodology): “, 2016), or TED, 23 languages, (Cettolo et al., 2012).”