1. Top-K Off-Policy Correction for a REINFORCE Recommender System, WSDM 2019, Google 作者...
1. Top-K Off-Policy Correction for a REINFORCE Recommender System, WSDM 2019, Google 作者...
https://github.com/tensorflow/ranking tf ranking 的代码里面有list wise loss 实现代码
End-to-End Query Term Weighting https://research.google/pubs/end-to-end-query-term-weig...
https://www.cnblogs.com/zhouyc/p/16943042.html https://zhuanlan.zhihu.com/p/250851222 h...
https://www.58pic.com/ 千图 https://www.bigbigwork.com/home 大作 https://huaban.com/guide 花...
1.链路一致性问题 2.冷启动问题 3.转化稀疏,长尾问题 4.延迟转化问题 5.短长期目标平衡问题 6.多目标平衡问题
论文Fresh Content Needs More Attention:Multi-funnel Fresh Content Recommendation 方案和系统都不复...
“深度”学习计算广告 https://baijiahao.baidu.com/s?id=1752993274919918339&wfr=spider&for=pc
爱奇艺 https://www.infoq.cn/article/lthcdazelzgc639p1p5q
推荐系统难题挑战(7):POSO,从模型角度解决用户冷启动问题 https://zhuanlan.zhihu.com/p/472726462 从零单排推荐系统[https:/...
阿里文娱智能营销增益模型 ( Uplift Model ) 技术实践 https://blog.51cto.com/u_15322177/5055434
点击率预测模型Embedding层的学习和训练 https://www.bilibili.com/read/cv16447478/
电商搜索实践,主要是架构 https://developer.aliyun.com/article/783767?spm=a2c6h.12873639.article-det...
Privileged Features Distillation at Taobao Recommendations 里面还有怎么在召回做distillation Slate...
前沿重器[28] | 前沿的向量召回都是怎么做的 https://mp.weixin.qq.com/s?__biz=MzIzMzYwNzY2NQ==&mid=22474883...
https://mp.weixin.qq.com/s/XETtP9IefIhK1vgcJtZNhQ https://mp.weixin.qq.com/s/svVl1eiVUH...
https://www.6aiq.com/article/1575265821104 多任务loss 优化https://blog.csdn.net/wuzhongqiang...
conda list 查看自己已安装包 conda create -n env_name(自己写名字) python=3.7(版本号) 创建虚拟环境 conda activ...