1.xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems
2. 廣告領(lǐng)域模型發(fā)展梳理以及最新模型xDeepFM解讀
3. xDeepFM詳解
4. xDeepFM github

comments on DCN:
CrossNet learns a special type of high-order feature interactions, where each hidden layer in the CrossNet is a scalar multiple of x0. However, note that the scalar multiple does not mean x_k is linear with x0. The CrossNet can learn feature interactions very efficiently (the complexity is negligible compared with a DNN model), however the downsides are: (1) the output of CrossNet is limited in a special form, with each hidden layer is a scalar multiple of x0; (2) interactions come in a bit-wise fashion.
DCN代碼實(shí)現(xiàn)

deepFM:
deepFM代碼實(shí)現(xiàn)