الباحثون

Zhaoxin Fan

المنشورات 2

نسخة أولية وصول مفتوح

EagleDepth: Efficient Fine-Grained Depth Estimation via Pixel Diffusion Decoder

Bowen Chai, Tianbao Zhang, Shuyu Wu وآخرون · 2026

Recovering detailed geometry from high-resolution images is critical for precise perception of the surroundings and objects. However, existing methods which use latent-space modeling and VAE reconstruction can compromise geometric details. Furthermore, decoding from latent codes introduces substantial inference overhea …

نسخة أولية وصول مفتوح

Cascade: Hierarchical Recoverability Control for Large Language Model Unlearning

Qingchen Yu, Shiying Duan, Xiaodong Li وآخرون · 2026

Large Language Model (LLM) unlearning is essential for removing sensitive or copyrighted knowledge while preserving general utility. Existing methods often leave residual knowledge in intermediate representations, which can still be recovered. To address this, we propose Cascade, a hierarchical recoverability control f …

المؤلفون المشاركون