DeepAI AI Chat
Log In Sign Up

DialectGram: Automatic Detection of Dialectal Variation at Multiple Geographic Resolutions

by   Hang Jiang, et al.
Stanford University

We propose DialectGram, a method to detect dialectical variation across multiple geographic resolutions. In contrast to prior work, which requires apriori knowledge of the geographic resolution and the set of regions, DialectGram automatically infers dialect-sensitive senses without these constraints using a nonparametric Bayesian extension of Skip-gram. Consequently, DialectGram only needs one-time training to enable an analysis of dialectical variation at multiple resolutions. To validate our approach, and establish a quantitative benchmark, we create a new corpus Geo-Tweets2019 with English tweets from the US and the UK, and new validation set DialectSim for evaluating word embeddings in American and British English.


DialectGram: Detecting Dialectal Variation at Multiple Geographic Resolutions

Several computational models have been developed to detect and analyze d...

Breaking Sticks and Ambiguities with Adaptive Skip-gram

Recently proposed Skip-gram model is a powerful method for learning high...

Compressing Word Embeddings Using Syllables

This work examines the possibility of using syllable embeddings, instead...

Predicting Demographics of High-Resolution Geographies with Geotagged Tweets

In this paper, we consider the problem of predicting demographics of geo...

Freshman or Fresher? Quantifying the Geographic Variation of Internet Language

We present a new computational technique to detect and analyze statistic...

A Computational Approach to Automatic Prediction of Drunk Texting

Alcohol abuse may lead to unsociable behavior such as crime, drunk drivi...

Semi-supervised Classification using Attention-based Regularization on Coarse-resolution Data

Many real-world phenomena are observed at multiple resolutions. Predicti...

Code Repositories


[SCiL 2020] DialectGram: Automatic Detection of Dialectal Changes with Multi-geographic Resolution Analysis

view repo