Cees G. M. Snoek

Full Professor@University of Amsterdam·荷兰·OpenReview

研究方向

Multimodal Foundation Models · Test-Time Adaptation & Generalization · Action & Event Recognition · Video Understanding · Multimodal Learning · Large-Scale Visual Retrieval

Cees G. M. Snoek

What Layers When: Learning to Skip Compute in LLMs with Residual Gates

MoAlign: Motion-Centric Representation Alignment for Video Diffusion Models

Structured-Noise Masked Modeling for Video, Audio and Beyond

NeoBabel: An Inclusive Multilingual Open Tower for Visual Generation

RegionReasoner: Region-Grounded Multi-Round Visual Reasoning

Prompt-Robust Vision-Language Models via Meta-Finetuning

Not All Tokens Matter Equally: Uncertainty-Guided Test Time Prompt Adaptation

Single-Sample Test-Time Reinforcement Learning for Vision-Language Models

3DPoV: Improving 3D understanding via Patch Ordering on Videos

KV Cache Steering for Controlling Frozen LLMs

Memory Forgetting Adapter Sculpting for Selective Multimodal Large Language Model Unlearning

Selective Seeing: Context-Aware Attention Interventions for Mitigating Hallucinations in Large Vision-Language Models

Purrception: Variational Flow Matching for Vector-Quantized Image Generation

QuantiShift: Any-Shift Object Quantification by Text-to-Image Diffusion Models

Visual Odometry with Transformers

Union-over-Intersections: Object Detection beyond Winner-Takes-All

One Hundred Neural Networks and Brains Watching Videos: Lessons from Alignment

TULIP: Token-length Upgraded CLIP

An Image is Worth More Than 16x16 Patches: Exploring Transformers on Individual Pixels

DynaPrompt: Dynamic Test-Time Prompt Tuning

CaPo: Cooperative Plan Optimization for Efficient Embodied Multi-Agent Cooperation

Elastic ViTs from Pretrained Models without Retraining

Geometric Neural Process Fields

TVBench: Redesigning Video-Language Evaluation

Prompt Diffusion Robustifies Any-Modality Prompt Learning

Learning to Ground VLMs without Forgetting