QMUL-SDS @ DIACR-Ita: Evaluating Unsupervised Diachronic Lexical Semantics Classification in Italian

TL;DR

This paper evaluates unsupervised methods for detecting semantic changes in Italian words over time, demonstrating that Temporal Word Embeddings with a Compass C-BOW model outperform other approaches with high accuracy.

Contribution

It introduces an effective unsupervised approach using Temporal Word Embeddings with a Compass C-BOW model for diachronic lexical semantics in Italian.

Findings

Temporal Word Embeddings with Compass C-BOW outperform other models

Achieved 83.3% accuracy, ranking 3rd in the task

Demonstrated effectiveness of unsupervised semantic change detection

Abstract

In this paper, we present the results and main findings of our system for the DIACR-ITA 2020 Task. Our system focuses on using variations of training sets and different semantic detection methods. The task involves training, aligning and predicting a word's vector change from two diachronic Italian corpora. We demonstrate that using Temporal Word Embeddings with a Compass C-BOW model is more effective compared to different approaches including Logistic Regression and a Feed Forward Neural Network using accuracy. Our model ranked 3rd with an accuracy of 83.3%.