
Muhammad Abdul-Mageed
machine translation
large language models
nlp
natural language generation
efficiency
chatgpt
llm
contrastive learning
arabic nlp
social media
language models
transfer learning
multilinguality
african languages
diacritization
42
presentations
26
number of views
4
citations
SHORT BIO
Muhammad Abdul-Mageed is a Canada Research Chair in Natural Language Processing and Machine Learning, and Associate Professor in the School of Information and Department of Linguistics (Joint Appointment), and Computer Science (Associate Member), at the University of British Columbia. Dr. Abdul-Mageed's research is in deep learning and natural language processing. His research program focuses on deep representation learning and natural language socio-pragmatics, with a goal to innovate more equitable, efficient, and 'social' machines for improved human health, safer social networking, and reduced information overload. Applications of his work currently span a wide range of speech and language understanding and generation tasks. For example, his group works on language models, automatic speech processing, machine translation, and computational socio-pragmatics in social media. Moreover, Dr. Abdul-Mageed directs the UBC Deep Learning & NLP Group, co-directs the SSHRC-funded I Trust AI Partnership Grant, and co-leads the SSHRC Ensuring Full Literacy Partnership Grant. He is a founding member of the Center for Artificial Intelligence Decision Making and Action and a member of the Institute for Computing, Information, and Cognitive Systems. He has published more than 120 research papers in peer-reviewed venues. His work has won more than half a dozen best paper awards and has been funded by Google, AMD, the Natural Sciences and Engineering Research Council of Canada (NSERC), the Social Sciences and Humanities Research Council of Canada (SSHRC), and the Canadian Foundation for Innovation (CFI), among others.
Presentations

Towards Zero-Shot Text-To-Speech for Arabic Dialects
Khai Doan and 2 other authors

John vs. Ahmed: Debate-Induced Bias in Multilingual LLMs
Anastasiia Demidova and 4 other authors

On the Utility of Pretraining Language Models on Synthetic Data
Alcides Alcoba Inciarte and 3 other authors

Benchmarking LLaMA-3 on Arabic Language Generation Tasks
Md Tawkat Islam Khondaker and 4 other authors

Dallah: A Dialect-Aware Multimodal Large Language Model for Arabic
Fakhraddin Alwajih and 2 other authors

NADI 2024: The Fifth Nuanced Arabic Dialect Identification Shared Task
AbdelRahim Elmadany and 7 other authors

WojoodNER 2024: The Second Arabic Named Entity Recognition Shared Task
Nagham Hamad and 5 other authors

Mixture-of-Supernets: Improving Weight-Sharing Supernet Training with Architecture-Routed Mixture-of-Experts
Ganesh Jawahar and 12 other authors

To Distill or Not to Distill? On the Robustness of Robust Knowledge Distillation
Abdul Waheed and 2 other authors

Toucan: Many-to-Many Translation for 150 African Language Pairs
AbdelRahim Elmadany and 2 other authors

FinTral: A Family of GPT-4 Level Multimodal Financial Large Language Models
Gagan Bhatia and 3 other authors

Peacock: A Family of Arabic Multimodal Large Language Models and Benchmarks
Fakhraddin Alwajih and 4 other authors

Cheetah: Natural Language Generation for 517 African Languages
Ife Adebara and 2 other authors

LLM Performance Predictors are good initializers for Architecture Search
Ganesh Jawahar and 3 other authors

Interplay of Machine Translation, Diacritics, and Diacritization
Wei-Rui Chen and 2 other authors

Fumbling in Babel: An Investigation into ChatGPT’s Language Identification Ability
Wei-Rui Chen and 4 other authors