# dlrm報告


## Embedding lookup
### Categorical data
Categorical data(類別型資料)是把input feature一個或多個不連續的項目變成一個集合，這個集合可以像是某個user看過哪些電影、這個文件出現過哪些字之類的。
在DLRM裡 會使用embedding table將sparse features映射成dense representations
> each sparse ID is used to look-up a unique row in embedding table
> Each vector is paried with an Embedding table using Sparse ID

## MLP
[MLP說明及推導](https://chih-sheng-huang821.medium.com/%E6%A9%9F%E5%99%A8%E5%AD%B8%E7%BF%92-%E7%A5%9E%E7%B6%93%E7%B6%B2%E8%B7%AF-%E5%A4%9A%E5%B1%A4%E6%84%9F%E7%9F%A5%E6%A9%9F-multilayer-perceptron-mlp-%E5%90%AB%E8%A9%B3%E7%B4%B0%E6%8E%A8%E5%B0%8E-ee4f3d5d1b41)

* FINAL OUTPUT(event probability) -> from the sigmoid function
* INPUT -> Sum of Dot Product(Embedding + processed dense feature)

###### tags: `DLRM`