Thursday, January 25, 2024 - 15:00 to 16:30
The School of Computer Science is pleased to present…
LADy: Latent Aspect Detection via Backtranslation Augmentation
MSc Thesis Defense by: Farinam Hemmati Zadeh
Date: Thursday, January 25th, 2024
Time: 3:00 pm – 4:00 pm
Location: Essex Hall Room 122
Abstract:
Within the context of review analytics, aspects are the features of products and services at which customers target their opinions and sentiments. Aspect detection helps product owners and service providers identify shortcomings and prioritize customers’ needs. Existing methods focus on detecting the surface form of an aspect falling short when aspects are latent in reviews, especially in an informal context like in social posts. In this research work, we propose data augmentation via natural language back translation to extract latent occurrences of aspects. We presume that back translation (1) can reveal latent aspects because they may not be commonly known in the target language and can be generated through back translation; (2) augments context-aware synonymous aspects from a target language to the original language, hence addressing the out-of-vocabulary issue; and (3) helps with the semantic disambiguation of polysemous words and collocations. Through our experiments on well-known aspect detection methods across several datasets of restaurant and laptop reviews, we demonstrate that review augmentation via back translation yields a steady performance boost in baselines. We further contribute LADy, a benchmark library to support the reproducibility of our research, which is publicly available at https://github.com/fani-lab/LADy.
Keywords: Review analysis, Aspect detection, Back translation augmentation
Thesis Committee:
Internal Reader: Dr. Robin Gras
External Reader: Dr. Tanja Collet-Najem, Department of Languages, Literatures and Cultures
Advisor: Dr. Hossein Fani
Chair: Dr. Majid Afshar Noghondari
