AE (auto-encoders) The schematic structure of an autoencoder is as follows: The encoder part of the network is used for encoding and sometimes even f ...
Gumbel-sofmax采样技巧
Posted on
以强化学习为例，假设网络输出的三维向量代表三个动作（前进、停留、后退）在下一步的收益，value=[-10,10,15]，那么下一步我们就会选择收益最大的动作（后退）继续执行，于是输出动作[0,0,1]。选择值最大的作为输出动作，这样做本身没问题，但是在网络中这种取法有个问题是不能计算梯度，也就不能 ...
GANs notebook
Posted on
|
In
deep learning
初识GAN 生成模型与判别模型 理解对抗网络，首先要了解生成模型和判别模型。判别模型比较好理解，就像分类一样，有一个判别界限，通过这个判别界限去区分样本。从概率角度分析就是获得样本x属于类别y的概率，是一个条件概率$P（y \mid x)$。而生成模型是需要在整个条件内去产生数据的分布，就像高斯分 ...
CnGAN Generative Adversarial Networks for Cross-network user preference generation for non-overlapped users
Posted on
|
Edited on
|
In
recommender systems
Intro cross-network Recommendation(are more robust against cold-start and data sparsity issues) problem However, despite the growing success of cross- ...
Deep Learning Based Recommender System A Survey and New Perspectives
Posted on
|
Edited on
|
In
recommender systems
This article aims to provide a comprehensive review of recent research efforts on deep learning-based recommender systems. Intro We conduct a compreh ...
RippleNet Propagating User Preferences on the Knowledge Graph for Recommender Systems
Posted on
|
Edited on
|
In
recommender systems
introduction CF-->sparsity, the cold start-->side information-->knowledge graph KG's strength KG introduces semantic relatedness among items ...
matrix factorization techniques for recommender systems
Posted on
|
Edited on
|
In
recommender systems
Three strength of MF (conclusion): combining good scalability with predictive accuracy. offering much flexibility for modeling various rela-life situ ...
推荐系统经典论文
Posted on
|
In
recommender systems
综述类: 1、Towards the Next Generation of Recommender Systems: A Survey of the State-of-the-Art and Possible Extensions。最经典的推荐算法综述 2、Collaborative Filteri ...