@7/21/2025
주제: 4D Gaussian Editing using video editing model
1.
InFusion: Inpainting 3D Gaussians via Learning Depth Completion from Diffusion Prior [arxiv 2024.04]
2.
Señorita-2M: A High-Quality Instruction-based Dataset for General Video Editing by Video Specialists [arxiv 2025.05]
발표자: 정성훈
주제: Understanding Mixture-of-Experts: A study in Efficient Scaling
1. Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer [ICLR 2017]
2. MoE++: Accelerating Mixture-of-Experts Methods with Zero-Computation Experts [ICLR 2025 Oral]
발표자: 진인환
주제 : Motion modeling
1. Generative Image Dynamics [CVPR 2024 Best Paper Award]
2. Motion-I2V: Consistent and Controllable Image-to-Video Generation with Explicit Motion Modeling[ACM SIGGRAPH 2024]
발표자: 문형주
@7/28/2025
주제 : Object change detection
1. The Change You Want to See (Now in 3D) [ICCV workshop 2023]
2. Improving Zero-Shot Object-Level Change Detection by Incorporating Visual Correspondence [WACV 2025]
발표자: 이우진
주제 : Visual Place Recognition
1. ETR: An Efficient Transformer for Re-ranking in Visual Place Recognition [WACV 2023]
2. R2Former: Unified Retrieval and Reranking Transformer for Place Recognition[CVPR 2023]
발표자: 우민수
@8/14/2025
주제 : Massive Activation in transformer architecture
1.
Vision Transformers Need Registers [ICLR 2024]
2.
Efficient Streaming Language Models with Attention Sinks [ICLR 2024]
발표자: 백창우
주제 : Vision Token Reduction in LVLMs
1.
[CLS] Attention is All You Need for Training-Free Visual Token Pruning: Make VLM Inference Faster
2.
VScan: Rethinking Visual Token Reduction for Efficient Large Vision-Language Models
발표자: 김소현
@8/18/2025
주제 : VLM/MLLM-based Agentic Systems: Methods and Applications
1. AN INTELLIGENT AGENTIC SYSTEM FOR COMPLEXIMAGE RESTORATION PROBLEMS [ICLR 2025]
2. RestoreAgent: Autonomous Image Restoration Agent via Multimodal Large Language Models [NeurIPS 2024]
발표자: 정다운
주제 : Object Detection via language modality
1. YOLO-World: Real-Time Open-Vocabulary Object Detection [CVPR 2024]
2. Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection [ECCV 2024]
발표자: 김태훈
@8/26/2025
주제: Matryoshka Representation Learning with Structured Supervision
1. Matryoshka Representation Learning [NeurIPS 2022]
2. Dimension-Specific Margins and Element-Wise Gradient Scaling for Enhanced Matryoshka Speaker Embedding [IEEE Access, 2025]
발표자: 박순찬
주제: 3D Cinemagraphy
1. LoopGaussian: Creating 3D Cinemagraph with Multi-view Images via Eulerian Motion Field [ACM MM 2024]
2. Optimizing 4D Gaussians for Dynamic Scene Video from Single Landscape Images [ICLR 2025]
발표자: 김소형