SIGTYP 2020 Shared Task: Prediction of Typological Features

Abstract

Typological knowledge bases (KBs) such as WALS (Dryer and Haspelmath, 2013) contain information about linguistic properties of the world's languages. They have been shown to be useful for downstream applications, including cross-lingual transfer learning and linguistic probing. A major drawback hampering broader adoption of typological KBs is that they are sparsely populated, in the sense that most languages only have annotations for some features, and skewed, in that few features have wide coverage. As typological features often correlate with one another, it is possible to predict them and thus automatically populate typological KBs, which is also the focus of this shared task. Overall, the task attracted 8 submissions from 5 teams, out of which the most successful methods make use of such feature correlations. However, our error analysis reveals that even the strongest submitted systems struggle with predicting feature values for languages where few features are known.

Predicting Typological Features in WALS using Language Embeddings and Conditional Probabilities: ÚFAL Submission to the SIGTYP 2020 Shared Task

“…The SIGTYP 2020 shared task (Bjerva et al, 2020) splits the WALS data into training, development and test portion. In the blind test data, some feature values are masked and the participating system is supposed to predict them based on the remaining features that are left visible.…”

KMI-Panlingua-IITKGP @SIGTYP2020: Exploring rules and hybrid systems for automatic prediction of typological features

“…The dataset 1 used for this experiment was extracted from World Atlas of Language Structures (WALS) (Dryer and Haspelmath, 2013). It covered the typological features of close to 2,000 languages (Bjerva et al, 2020). These typological features were organised in 8 columns (including Language ID, Language name, Latitude, Longitude, Genus, Family, Country Codes, and feature-value.…”

NEMO: Frequentist Inference Approach to Constrained Linguistic Typology Feature Prediction in SIGTYP 2020 Shared Task

“…Developing method ologies for accurately predicting missing typolog ical features on the basis of existing knowledge is therefore crucial for a wider adoption of typologi cal resources in NLP tasks and beyond (Evans and Levinson, 2009). This paper presents the work done by the "NEMO Team" (Google London and Tokyo) on the constrained subtask for the SIGTYP 2020 Shared Task (Bjerva et al, 2020). We experimented with a variety of machine learning models using only the features provided in the training, devel opment and test sets.…”

“…

This paper describes the NEMO submission to SIGTYP 2020 shared task (Bjerva et al, 2020) which deals with prediction of linguis tic typological features for multiple languages using the data derived from World Atlas of Language Structures (WALS). We employ fre quentist inference to represent correlations be tween typological features and use this repre sentation to train simple multiclass estimators that predict individual features.

…”