최근 딥러닝 기반 영상 생성 모델은 놀라운 발전을 이루고 있습니다. 이 발표에서는 특히 주목받는 Diffusion 모델에 초점을 맞춰 이론적 배경, CLIP 등 Visual Language Model 기반 프롬프트 활용, zero-shot 개인화 기술 등을 심층적으로 살펴보겠습니다.

Secure model aggregation is a key component of federated learning (FL) that aims at protecting the privacy of each user’s individual model while allowing for their global aggregation. State-of-the-art secure aggregation protocols rely on secret sharing of the random-seeds used for mask generations at the users. The complexity of such approaches, however, grows substantially with the number of participating users, which is a main bottleneck of scaling to large number of users. In this talk, I will introduce a new secure aggregation protocol, named LighteSecAgg, which overcomes the bottleneck as well as can be extended to various scenarios such as asynchronous FL and verifiable FL.

In this talk, I will present our latest research on developing generative models for creating highly realistic 3D digital humans. Three state-of-the-art approaches will be introduced: NCHO (ICCV 2023) for learning neural 3D composition of humans and clothing, Chupa (ICCV 2023) for creating 3D clothed humans using 2D diffusion probabilistic models, and GALA (CVPR 2024) for generating animatable layered assets from a single 3D scan. Key challenges, methodologies, and results will be discussed, along with insights into the broader impacts and future directions of this field.