paper阅读
czlm爱你的笑
努力
展开
-
标签推荐领域阅读
标签推荐2006-Information retrieval in folksonomies: Search and ranking引用数:1239 来源:European semantic web conferencetakes into account the folksonomy structure for ranking search requests in folksonomy based systems. The algorithm will be used for two purpose原创 2020-11-06 22:47:52 · 609 阅读 · 0 评论 -
推荐领域论文阅读
推荐2001-Item-based collaborative filtering recommendation algorithms引用数:9183 来源:Proceedings of the 10th international conference on World Wide Web问题:访问人数越来越多。传统基于user-CF方法的工作量随着用户的增多而增大,通过查找相似用户来进行推荐必然会碰到瓶颈。CF模型将整个用户-物品m×nm\times nm×n评分作为输入矩阵,可分为基于memo原创 2020-11-06 19:10:17 · 511 阅读 · 0 评论 -
Fashion领域论文阅读
Fashion推荐2018-Interpretable multimodal retrieval for fashion products引用数:23 来源:Proceedings of the 26th ACM international conference on Multimedia.2019-Complete the look: Scene-based complementary product recommendation引用数:20 来源:Proceedings of the IEEE原创 2020-11-06 19:09:44 · 990 阅读 · 0 评论 -
神经网络论文阅读
卷积CNN2014-Very deep convolutional networks for large-scale image recognition引用数:44780 来源:ICLR( International Conference on Learning Representations)the effect of the convolutional network depth on its accuracy,using an architecture with very small (3×3原创 2020-11-06 19:09:14 · 410 阅读 · 0 评论 -
Towards Deeper Graph Neural Networks
问题:一层的GCN layer只考虑直接邻域,加深模型层数可获得更大的感受野receptive fields,但性能却会变坏。以前认为是over-smoothing问题导致的。本文增加层性能变坏是因为GCN模型里两个关键部分——representation transformation和propagation——的纠缠导致的。分离这两个部分可以解决该问题。模型:进行特征变换,堆叠层,只进行propagaition,不进行特征变换。adaptive adjustment,因为层数不原创 2020-11-05 19:24:20 · 1202 阅读 · 0 评论