Joseph Marvin Imperial is a UKRI CDT Doctoral Researcher at the University of Bath's Integrated Ph.D. Program in Accountable, Responsible, and Transparent AI (also called ART-AI). He conducts research in NLP and Machine Learning, particularly in the following areas:

1. Controlled Natural Language Generation (Standardize Project, EMNLP2023, GEM2023).

2. Building Low-Resource Assessment Models (SEACrowd Project, EMNLP2023, ACL2023, BEA2022, AIED2022).

3. Building Low-Resource Language Corpora (PACLIC 2023, EMNLP2023, NAACL2024).

4. Benchmark Safety and Risks of Gen AI Models (ICML2024, IEEE Spectrum, MLCommons Whitepaper).

He is affiliated with the BathNLP group and is advised by Dr. Harish Tayyar Madabushi and Dr. Gail Forey. Joseph has been awarded the Doctoral Recognition Award in 2024 for his contributions to NLP research of the university.

Joseph is a proactive member of the NLP community and has served as a reviewer for top conferences such as NeurIPS, ICLR, ICML, AAAI, and *ACL conferences and workshops—recently winning Outstanding Reviewer awards at ICML 2024 and ACL 2023. He has been involved with numerous projects, such as community-led initiatives like the Aya Multilingual Project by Cohere for AI and government-funded research projects such as CHED-LAKAS (Leading the Advancement of Knowledge in Agriculture and Sciences). He is also a Regular Member of the National Research Council of the Philippines under the Mathematical Sciences Division.

Outside of Bath, Joseph is affiliated with the Department of Computer Science and the Human Language Technology (HLT) Lab at National University Philippines, but is currently on study leave.