← Back to papers
2022 SEMEVAL SemEval 2022

ISCAS at SemEval-2022 Task 10: An Extraction-Validation Pipeline for Structured Sentiment Analysis

Abstract

AbstractISCAS participated in both sub-tasks in SemEval-2022 Task 10: Structured Sentiment competition. We design an extraction-validation pipeline architecture to tackle both monolingual and cross-lingual sub-tasks. Experimental results show the multilingual effectiveness and cross-lingual robustness of our system. Our system is openly released on: https://github.com/luxinyu1/SemEval2022-Task10/.

🧭 Keyword Pioneer — validation pipeline
🐝 Cross-Pollinator — Artificial Intelligence, Deep Learning, Interdisciplinary, Machine Learning, Natural Language Processing

Authors

Xinyu Lu , Mengjie Ren , Yaojie Lu , Hongyu Lin

Topics

Natural Language Processing > Understanding > Sentiment Analysis Natural Language Processing > Resources & Methods > Multilingual NLP

Keywords

sentiment extraction validation pipeline cross-lingual robustness multilingual effectiveness
Download PDF

Related papers

YMAI at SemEval-2022 Task 5: Detecting Misogyny in Memes using VisualBERT and MMBT MultiModal Pre-trained Models 2022
MS@IW at SemEval-2022 Task 4: Patronising and Condescending Language Detection with Synthetically Generated Data 2022
CS-UM6P at SemEval-2022 Task 6: Transformer-based Models for Intended Sarcasm Detection in English and Arabic 2022
HW-TSC at SemEval-2022 Task 7: Ensemble Model Based on Pretrained Models for Identifying Plausible Clarifications 2022
HITMI&T at SemEval-2022 Task 4: Investigating Task-Adaptive Pretraining And Attention Mechanism On PCL Detection 2022