Arpit Garg

Arpit Garg is a Research Fellow at the Australian Institute for Machine Learning (AIML), Faculty of Sciences, Engineering and Technology, University of Adelaide. His research develops robust and responsible AI systems for real-world environments with imperfect data, specializing in instance-dependent noisy label learning and multimodal large language models for Trust & Safety applications. He currently holds a Grant-Funded Research Fellow position at the Responsible AI Research Centre.

His primary research spans Artificial Intelligence, Machine Learning, and Computer Vision with strong focus on Trustworthy AI systems. He pioneered graphical model approaches for noisy label learning and advanced MLLMs through vision-language model integration, achieving significant AUC improvements in Trust & Safety applications. His work bridges theoretical innovation with high-impact implementations including film productions (Mad Max: Furiosa, Wolverine & Deadpool) and revenue-generating AI systems.

Scientific Awards: No specific awards were documented in the source materials.

Dr Garg serves as eligible Co-Supervisor for Masters and PhD candidates. His research is supported by active grants at the Responsible AI Research Centre, enabling collaborations with international institutions and regular contributions to top-tier conferences (ECCV, WACV, ICCV). Current projects focus on privacy-preserving frameworks for Australian data sovereignty and AI safety protocols for critical infrastructure.

He operates within the Australian Institute for Machine Learning ecosystem at Lot 14 campus (Room G.05.01), contributing to the University's strategic AI initiatives while maintaining industry partnerships including prior work at TikTok.