We present the HPLT (High Performance Language Technologies) language resources, a new massive multilingual dataset including both monolingual and bilingual corpora extracted from CommonCrawl and previously unused web crawls from the Internet Archive. We describe our methods for data acquisition, management and processing of large corpora, which rely on open-source software tools and high-performance computing. Our monolingual collection focuses on low- to medium-resourced languages and covers 75 languages and a total of ≈ 5.6 trillion word tokens de-duplicated on the document level. Our English-centric parallel corpus is derived from its monolingual counterpart and covers 18 language pairs and more than 96 million aligned sentence pairs with roughly 1.4 billion English tokens. The HPLT language resources are one of the largest open text corpora ever released, providing a great resource for language modeling and machine translation training. We publicly release the corpora, the software, and the tools used in this work.

Language identification is a crucial component in the automated production of language resources, particularly in multilingual and big data contexts. However, commonly used language identifiers struggle to differentiate between similar or closely-related languages. This paper introduces FastSpell, a language identifier that combines fastText (a pre-trained language identifier tool) and Hunspell (a spell checker) with the aim of having a refined second-opinion before deciding which language should be assigned to a text. We provide a description of the FastSpell algorithm along with an explanation on how to use and configure it. To that end, we motivate the need of such a tool and present a benchmark including some popular language identifiers evaluated during the development of FastSpell. We show how FastSpell is useful not only to improve identification of similar languages, but also to identify new ones ignored by other tools.

We present the most relevant results of the project MaCoCu: Massive collection and curation of monolingual and bilingual data: focus on under-resourced languages in its second year. To date, parallel and monolingual corpora have been produced for seven low-resourced European languages by crawling large amounts of textual data from selected top-level domains of the Internet; both human and automatic evaluation show its usefulness. In addition, several large language models pretrained on MaCoCu data have been published, as well as the code used to collect and curate the data.

Quality assessment has been an ongoing activity of the series of ParaCrawl efforts to crawl massive amounts of parallel data from multilingual websites for 29 languages. The goal of ParaCrawl is to get parallel data that is good for machine translation. To prove so, both, automatic (extrinsic) and human (intrinsic and extrinsic) evaluation tasks have been included as part of the quality assessment activity of the project. We sum up the various methods followed to address these evaluation tasks for the web-crawled corpora produced and their results. We review their advantages and disadvantages for the final goal of the ParaCrawl project and the related ongoing project MaCoCu.

This paper describes the experiments carried out during the development of the latest version of Bicleaner, named Bicleaner AI, a tool that aims at detecting noisy sentences in parallel corpora. The tool, which now implements a new neural classifier, uses state-of-the-art techniques based on pre-trained transformer-based language models fine-tuned on a binary classification task. After that, parallel corpus filtering is performed, discarding the sentences that have lower probability of being mutual translations. Our experiments, based on the training of neural machine translation (NMT) with corpora filtered using Bicleaner AI for two different scenarios, show significant improvements in translation quality compared to the previous version of the tool which implemented a classifier based on Extremely Randomized Trees.

We introduce the project “MaCoCu: Massive collection and curation of monolingual and bilingual data: focus on under-resourced languages”, funded by the Connecting Europe Facility, which is aimed at building monolingual and parallel corpora for under-resourced European languages. The approach followed consists of crawling large amounts of textual data from carefully selected top-level domains of the Internet, and then applying a curation and enrichment pipeline. In addition to corpora, the project will release successive versions of the free/open-source web crawling and curation software used.

We report on methods to create the largest publicly available parallel corpora by crawling the web, using open source software. We empirically compare alternative methods and publish benchmark data sets for sentence alignment and sentence pair filtering. We also describe the parallel corpora released and evaluate their quality and their usefulness to create machine translation systems.

This paper shows the utility of two open-source tools designed for parallel data cleaning: Bifixer and Bicleaner. Already used to clean highly noisy parallel content from crawled multilingual websites, we evaluate their performance in a different scenario: cleaning publicly available corpora commonly used to train machine translation systems. We choose four English–Portuguese corpora which we plan to use internally to compute paraphrases at a later stage. We clean the four corpora using both tools, which are described in detail, and analyse the effect of some of the cleaning steps on them. We then compare machine translation training times and quality before and after cleaning these corpora, showing a positive impact particularly for the noisiest ones.

This paper describes Prompsit Language Engineering’s submissions to the WMT 2018 parallel corpus filtering shared task. Our four submissions were based on an automatic classifier for identifying pairs of sentences that are mutual translations. A set of hand-crafted hard rules for discarding sentences with evident flaws were applied before the classifier. We explored different strategies for achieving a training corpus with diverse vocabulary and fluent sentences: language model scoring, an active-learning-inspired data selection algorithm and n-gram saturation. Our submissions were very competitive in comparison with other participants on the 100 million word training corpus.