HKESG at the ML-ESG Task: Exploring Transformer Representations for Multilingual ESG Issue Identification

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

Abstract

Environmental, Social and Governance reports have to be periodically released by financial companies, as they represent an essential guide for the potential, socially-responsible new investors. Therefore, automatizing the analysis of reports and extracting the main ESG issues mentioned in the text is a goal of primary importance for financial Natural Language Processing (NLP) systems.

In this paper, we report our experiments for the FinSim4-ESG Shared Task, dedicated to the problem of multilingual ESG issue identification in English and French. Our results show that even simple classifiers trained on multilingual data and using crosslingual Transformer representations can achieve a strong performance in the task.
Original languageEnglish
Title of host publicationProceedings of the IJCAI Joint Workshop on Financial Technology and Natural Language Processing (FinNLP) and Multimodal AI for Financial Forecasting (MUFFIN)
PublisherAssociation for Computational Linguistics (ACL)
Publication statusPublished - Jul 2023
EventThe Joint Workshop of the 5th Financial Technology and Natural Language Processing (FinNLP) and 2nd Multimodal AI For Financial Forecasting (MUFFIN) - Sheraton Grand Macau, Macau, China
Duration: 20 Aug 2023 → …
https://finnlp-muffin-ijcai23.github.io/

Conference

ConferenceThe Joint Workshop of the 5th Financial Technology and Natural Language Processing (FinNLP) and 2nd Multimodal AI For Financial Forecasting (MUFFIN)
Abbreviated titleFinNLP 2023
Country/TerritoryChina
CityMacau
Period20/08/23 → …
Internet address

Fingerprint

Dive into the research topics of 'HKESG at the ML-ESG Task: Exploring Transformer Representations for Multilingual ESG Issue Identification'. Together they form a unique fingerprint.

Cite this