diff --git a/README.md b/README.md index 9c826a091f24659b5f7dd2f3de9658c14fe58f59..eeec2145a98c28088b5e7c26503c204af8c69c4d 100644 --- a/README.md +++ b/README.md @@ -59,7 +59,8 @@ | 排序 | [xDeepFM](models/rank/xdeepfm/model.py) | ✓ | x | ✓ | x | [KDD 2018][xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems](https://dl.acm.org/doi/pdf/10.1145/3219819.3220023) | | 排序 | [DIN](models/rank/din/model.py) | ✓ | x | ✓ | x | [KDD 2018][Deep Interest Network for Click-Through Rate Prediction](https://dl.acm.org/doi/pdf/10.1145/3219819.3219823) | | 排序 | [Wide&Deep](models/rank/wide_deep/model.py) | ✓ | x | ✓ | x | [DLRS 2016][Wide & Deep Learning for Recommender Systems](https://dl.acm.org/doi/pdf/10.1145/2988450.2988454) | - | 排序 | [FGCNN](models/rank/fgcnn/model.py) | ✓ | ✓ | ✓ | ✓ | [WWW 2019][Feature Generation by Convolutional Neural Network for Click-Through Rate Prediction](https://arxiv.org/pdf/1904.04447.pdf) | + | 排序 | [FGCNN](models/rank/fgcnn/model.py) | ✓ | ✓ | ✓ | ✓ | [WWW 2019][Feature Generation by Convolutional Neural Network for Click-Through Rate Prediction](https://arxiv.org/pdf/1904.04447.pdf) + | 排序 | [Fibinet](models/rank/fibinet/model.py) | ✓ | ✓ | ✓ | ✓ | [RecSys19][FiBiNET: Combining Feature Importance and Bilinear feature Interaction for Click-Through Rate Prediction]( https://arxiv.org/pdf/1905.09433.pdf) | | 多任务 | [ESMM](models/multitask/esmm/model.py) | ✓ | ✓ | ✓ | ✓ | [SIGIR 2018][Entire Space Multi-Task Model: An Effective Approach for Estimating Post-Click Conversion Rate](https://arxiv.org/abs/1804.07931) | | 多任务 | [MMOE](models/multitask/mmoe/model.py) | ✓ | ✓ | ✓ | ✓ | [KDD 2018][Modeling Task Relationships in Multi-task Learning with Multi-gate Mixture-of-Experts](https://dl.acm.org/doi/abs/10.1145/3219819.3220007) | | 多任务 | [ShareBottom](models/multitask/share-bottom/model.py) | ✓ | ✓ | ✓ | ✓ | [1998][Multitask learning](http://reports-archive.adm.cs.cmu.edu/anon/1997/CMU-CS-97-203.pdf) | diff --git a/doc/imgs/fibinet.png b/doc/imgs/fibinet.png new file mode 100644 index 0000000000000000000000000000000000000000..71d9b6c14219cdbf55b45d206206d11da6dc500a Binary files /dev/null and b/doc/imgs/fibinet.png differ diff --git a/models/rank/readme.md b/models/rank/readme.md old mode 100755 new mode 100644 index 51438fd65c5d36c351815ab903b9864db3cdd2c1..3ca96de21b7e134270adc448107c9db59f3632c3 --- a/models/rank/readme.md +++ b/models/rank/readme.md @@ -37,35 +37,43 @@ | xDeepFM | xDeepFM | [xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems](https://dl.acm.org/doi/pdf/10.1145/3219819.3220023)(2018) | | DIN | Deep Interest Network | [Deep Interest Network for Click-Through Rate Prediction](https://dl.acm.org/doi/pdf/10.1145/3219819.3219823)(2018) | | FGCNN | Feature Generation by CNN | [Feature Generation by Convolutional Neural Network for Click-Through Rate Prediction](https://arxiv.org/pdf/1904.04447.pdf)(2019) | +| FIBINET | Combining Feature Importance and Bilinear feature Interaction | [《FiBiNET: Combining Feature Importance and Bilinear feature Interaction for Click-Through Rate Prediction》]( https://arxiv.org/pdf/1905.09433.pdf) | 下面是每个模型的简介（注：图片引用自链接中的论文） [wide&deep](https://dl.acm.org/doi/pdf/10.1145/2988450.2988454): +

- [DeepFM](https://arxiv.org/pdf/1703.04247.pdf): +

- [XDeepFM](https://dl.acm.org/doi/pdf/10.1145/3219819.3220023): +

- [DCN](https://dl.acm.org/doi/pdf/10.1145/3124749.3124754): +

- [DIN](https://dl.acm.org/doi/pdf/10.1145/3219819.3219823): +

+[FIBINET](https://arxiv.org/pdf/1905.09433.pdf): + +

+ +

## 使用教程(快速开始) + 使用样例数据快速开始，参考[训练](###训练) & [预测](###预测) ## 使用教程(复现论文) 为了方便使用者能够快速的跑通每一个模型，我们在每个模型下都提供了样例数据，并且调整了batch_size等超参以便在样例数据上更加友好的显示训练&测试日志。如果需要复现readme中的效果请按照如下表格调整batch_size等超参，并使用提供的脚本下载对应数据集以及数据预处理。