Domain Adaptation on Wheels: Closing the Gap to the Open-world

Explore domain adaptation techniques for autonomous driving applications in this comprehensive seminar. Delve into the challenges of deep models failing when input data shifts away from their operational design domain. Learn about unsupervised domain adaptation for semantic segmentation, including approaches like AdvEnt, DADA, xMUDA, and ConDA. Discover practical adaptations of domain adaptation through BUDA, MTAF, and MuHDi techniques. Examine the latest developments in adapting to unseen scenarios with the PODA method. Gain insights from research scientist Tuan-Hung Vu on label-efficient learning of robust and reliable perception models, covering topics such as domain adaptation, domain generalization, robustness, zero-shot learning, and open-vocabulary learning.

[Seminar Series] Domain Adaptation on Wheels: Closing the Gap to the Open-world