INCUR
发表于 2025-3-23 13:38:44
,Meta-Learning with Less Forgetting on Large-Scale Non-Stationary Task Distributions,tributions sequentially arrive with some ORDER), to tackle these two major challenges. Specifically, our ORDER introduces a novel mutual information regularization to robustify the model with unlabeled OOD data and adopts an optimal transport regularization to remember previously learned knowledge i
crockery
发表于 2025-3-23 15:04:41
,DnA: Improving Few-Shot Transfer Learning with Low-Rank Decomposition and Alignment,the low-rank subspace), and the extra flexibility to absorb the new out-of-the-domain knowledge (via freeing the sparse residual). Our resultant framework, termed Decomposition-and-Alignment (.), significantly improves the few-shot transfer performance of the SS pre-trained model to downstream tasks
打包
发表于 2025-3-23 18:22:27
http://reply.papertrans.cn/24/2343/234275/234275_13.png
无效
发表于 2025-3-23 22:10:36
,Open-World Semantic Segmentation via Contrasting and Clustering Vision-Language Embedding,mage encoder is jointly trained with a vision-based contrasting and a cross-modal contrasting, which encourage the visual embeddings to preserve both fine-grained semantics and high-level category information that are crucial for the segmentation task. Furthermore, an online clustering head is devis
Essential
发表于 2025-3-24 05:49:11
http://reply.papertrans.cn/24/2343/234275/234275_15.png
火海
发表于 2025-3-24 07:34:52
http://reply.papertrans.cn/24/2343/234275/234275_16.png
OTTER
发表于 2025-3-24 12:16:29
http://reply.papertrans.cn/24/2343/234275/234275_17.png
aristocracy
发表于 2025-3-24 18:04:44
http://reply.papertrans.cn/24/2343/234275/234275_18.png
Servile
发表于 2025-3-24 19:07:34
http://reply.papertrans.cn/24/2343/234275/234275_19.png
易碎
发表于 2025-3-25 02:13:32
http://reply.papertrans.cn/24/2343/234275/234275_20.png