Machine Learning for Detecting Hate Speech in Low Resource Languages
Year: 2020
Type: dissertation
Abstract: This work examines the role of both cross-lingual zero-shot learning and data augmentation\nin detecting hate speech online for low resource set-ups. The proposed\nsolutions for situations where the amount of labeled data is scarce are to use a\nlanguage with more resources during training or to create synthetic data points.\nCross-lingual zero-shot results suggest some knowledge transfer is occurring. However,\nresults seem greatly influenced by the specific training data set selected. This\nis further supported by cross-data set experimentation within the same language,\nwhere results were also found to fluctuate based on training data without the need\nfor cross-lingual transfer. Meanwhile, data augmentation methods show an improvement,\nespecially for low amounts of data. Furthermore, a detailed discussion\non how the proposed data augmentation techniques impact the data is presented in\nthis work.
Authors: David Rodríguez‐Rodríguez, Denitsa Saynova
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
SDG: Reduced inequalities
Open Access status: green
OpenAlex ID: W3166305028