A paper by Prof. Pascale FUNG from the Department of Electronic and Computer Engineering (ECE), her postgraduate students and collaborators from Salesforce Research won an Outstanding Paper Award at the 57th Annual Meeting of the Association for Computational Linguistics (ACL), held in Florence, Italy on July 28 – Aug 2, 2019. It is a highly competitive award with only five out of 3,800 paper submissions receiving it.

Entitled “Transferable Multi-Domain State Generator for Task-Oriented Dialogue Systems”, the paper proposes a transferable dialogue state generator (TRADE) to solve practical yet less studied problems in dialogue state tracking. It was co-authored by 2019 ECE MPhil graduate Wu Chien-Sheng, ECE PhD student Andrea Madotto, Prof. Fung, and Salesforce Research collaborators Ehsan Hosseini-Asl, Xiong Caiming and Richard Socher.

ACL is the premier conference of the field of computational linguistics, covering a broad spectrum of diverse research areas that are concerned with computational approaches to natural language. Prof. Fung was a keynote speaker at this year’s event to share on the topic of “Loquentes Machinae: Technical Approaches, Applications and Ethical Issues of Conversational Systems”.

In her talk, she gave an overview of some of the technical challenges, approaches and applications of conversational systems, and the debates on ethical issues surrounding them. She also highlighted some of the cultural differences in this area and discussed how people can collaborate internationally to build conversational systems that are secure, safe, and fair for all.

In addition, another paper co-authored by Prof. Fung and two of her PhD students was selected as an Outstanding Paper and received a Best Paper Award at the 4th Workshop on Representation Learning for NLP (RepL4NLP), which was hosted by ACL 2019. The paper was entitled “Learning Multilingual Meta-Embeddings for Code-Switching Named Entity Recognition” and the two students were Genta Indra Winata and Lin Zhaojiang, both from ECE.

The workshop was introduced as a synthesis of several years of independent computational linguistics workshops focusing on vector space models of meaning, compositionality, and the application of deep neural networks and spectral methods to NLP. It provides a forum for discussing recent advances on these topics, as well as future research directions in linguistically motivated vector-based models in NLP.