URIEL and lang2vec: Representing languages as typological, geographical, and phylogenetic vectors

By Lori S. Levin, Patrick Littell, David R. Mortensen, Ke Lin, Katherine Kairis, Carlisle Turner. Conference of the European Chapter of the Association for Computational Linguistics, 2017

304 citations recorded by the source corpus.

We introduce the URIEL knowledge base for massively multilingual NLP and the lang2vec utility, which provides information-rich vector identifications of languages drawn from typological, geographical, and phylogenetic databases and normalized to have straightforward and consistent formats, naming, and semantics. The goal of URIEL and lang2vec is to enable multilingual NLP, especially on less-resourced languages and make possible types of experiments (especially but not exclusively related to NLP tasks) that are otherwise difficult or impossible due to the sparsity and incommensurability of the data sources. lang2vec vectors have been shown to reduce perplexity in multilingual language modeling, when compared to one-hot language identification vectors.

Subfields: Artificial Intelligence, Cultural Studies

View Paper: https://www.semanticscholar.org/paper/e3a824e59ac66ceaaa420e0de4f2b5d809231f6e

How researchers see this paper

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

In Artificial Intelligence

In Cultural Studies

Citations in the corpus

Massively Multilingual Adversarial Speech Recognition (2019, methodology): “In order to determine language similarity in a principled way we used URIEL and lang2vec (Littell et al., 2017) to produce feature vectors for each language based on information from several linguistic resources before calculating their cosine similarity.”
MAD-G: Multilingual Adapter Generation for Efficient Cross-Lingual Transfer (2021, methodology): “To this end, Ponti et al. (2019b) exploit typological language vectors from the URIEL database (Littell et al., 2017) directly as language embeddings to generate a full set of model parameters.”
It’s not a Non-Issue: Negation as a Source of Error in Machine Translation (2020, methodology): “Brieﬂy, we obtain feature vectors for each language from lang2vec (Littell et al., 2017), and deﬁne the similarity between two languages as the cosine similarity between their feature vectors.”
From Zero to Hero: On the Limitations of Zero-Shot Language Transfer with Multilingual Transformers (2020, methodology): “…a simple regression on available transfer results, we show that we can (roughly) predict the transfer performance from (1) language proximity (Littell et al., 2017) for low-level tasks; (2) combination of language proximity and size of target-language pretraining corpora for high-level tasks.”
IGLUE: A Benchmark for Transfer Learning across Modalities, Tasks, and Languages (2022, methodology): “The latter is calculated as the cosine similarity between vectors of morpho-syntactic features extracted from WALS (Dryer & Haspelmath, 2013) by Littell et al. (2017).”
Choosing Transfer Languages for Cross-Lingual Learning (2019, methodology): “Speciﬁcally, we leverage six different linguistic distances queried from the URIEL Typological Database (Littell et al., 2017): The orthodromic distance between the languages on the surface of the earth, divided by the antipodal distance, based primarily on language location descriptions in Glottolo”
Simple and Effective Paraphrastic Similarity from Parallel Translations (2019, background): “The plot, shown in Figure 1, shows that sentencepiece (SP) overlap is highly corre-10 Table 5: Spearman’s ρ × 100 between average performance on the 2012-2016 STS tasks compared to SP overlap (SP Ovl.) and language distance as deﬁned by Littell et al. (2017).”
XTREME-R: Towards More Challenging and Nuanced Multilingual Evaluation (2021, background): “The typology index is based on the mean entropy of the distribution over 103 typological 10For simplicity, we treat Creole as a distinct language family (Bakker et al., 2011). features from URIEL (Littell et al., 2017) across the languages while the family index consists of the number of distinct la”
XCOPA: A Multilingual Dataset for Causal Commonsense Reasoning (2020, methodology): “(1993) and Miestamo (2004), we propose a series of simple and interpretable metrics that quantify diversity of a language sample independent of its size: 1) a typology index based on 103 typological features of each language from URIEL (Littell et al., 2017), originally sourced from the World Atlas ”
UDapter: Language Adaptation for Truly Universal Dependency Parsing (2020, methodology): “…we propose not to learn language embeddings from scratch, but to leverage a mix of linguistically curated and predicted typological features as obtained from the URIEL language typology database (Littell et al., 2017) which supports 3718 languages including all languages represented in the UD.”
Should All Cross-Lingual Embeddings Speak English? (2019, methodology): “The typological distance ( d gen ) between two languages can be approximated through their genealogical distance over hypothesized language family trees, which we obtain from the URIEL typo-logical database (Littell et al., 2017).”
MaSS: A Large and Clean Multilingual Corpus of Sentence-aligned Spoken Utterances Extracted from the Bible (2019, methodology): “Following this experiment, we investigated the correlation between the median rank and two variables: the quality of the alignment (human evaluation) and the syntactic distance between the language pairs (using the lang2vec library [31]).”
Learning Language Representations for Typology Prediction (2017, methodology): “We calculate these feature vectors using an NMT model trained on 1017 languages, and use them for typlogy prediction both on their own and in composite with feature vectors from previous work based on the genetic and geographic distance between languages (Littell et al., 2017).”
Multilingual Neural Machine Translation with Language Clustering (2019, background): “, 2019), such as language family, language typology or other language characteristics from the URIEL database (Levin et al., 2017).”
Ranking Transfer Languages with Pragmatically-Motivated Features for Multilingual Sentiment Analysis (2020, methodology): “tures are various types of distance between a pair of languages queried from the URIEL package (Littell et al., 2017): geographic (geo), genetic (gen), inventory (inv), syntactic (syn), phonological (phon) and featural (feat), adopted”
What Drives Cross-lingual Ranking? Retrieval Approaches with Multilingual Language Models (2025, unknown): “To assess whether retrieval performance correlates with linguistic proximity, we computed cosine similarity between query and document languages using typological vectors from lang2vec [76].”