Giter Site home page Giter Site logo

Comments (8)

louiswng avatar louiswng commented on August 21, 2024

这个公式为 hinge loss,优化时会拉近两个 z 之间的距离,也就是拉近两个用户间的交互相似度和社交相似度,从而消除社交视图的噪声。

from dsl.

Socialrec avatar Socialrec commented on August 21, 2024

如果两个用户之间交互相似度很低(即没有公共交互偏好),要使得该损失函数减小,不应该对应社交相似度变高吗?这不是矛盾了吗?抱歉,我没有理解。

from dsl.

Socialrec avatar Socialrec commented on August 21, 2024

因为您加了sigmoid函数,因此用户之间相似度肯定是正数。

from dsl.

louiswng avatar louiswng commented on August 21, 2024

在本文中,\hat z 与 z 是两个向量,他们的乘积表示这两个向量的相似度。要使得损失函数变小,则 \hat z 与 z 乘积对应增大,表示这两个向量的相似度变大,相当于拉近两个用户间的交互相似度和社交相似度。

from dsl.

Socialrec avatar Socialrec commented on August 21, 2024

但是您论文中和代码实现\hat z 与 z 都是一个数字🤔,如果是数字的话lable 是-1和1才能用hinge loss 吧。

from dsl.

louiswng avatar louiswng commented on August 21, 2024
  1. 不管用户间交互相似度的大小如何,要使得该损失函数减小,对应的社交相似度都会变高。但当某对用户间交互相似度低的时候,其社交相似度的优化梯度会较低;反之,社交相似度的优化梯度会较高。是通过这种方式来进行优化的。
  2. 在函数图像上,BPR Loss 比 Hinge Loss 更加平滑。我们也尝试过 BPR Loss,但实验效果较差。

from dsl.

Socialrec avatar Socialrec commented on August 21, 2024

“不管用户间交互相似度的大小如何,要使得该损失函数减小,对应的社交相似度都会变高”,按照这个说法,当用户间没有相似的兴趣偏好时,交互相似度小,社交相似度高,这本身就是矛盾的啊?和优化梯度大小有什么关系呢?

from dsl.

louiswng avatar louiswng commented on August 21, 2024

优化梯度可以从对变量求偏导理解。要使得该损失函数减小,当 z 很小时,\hat z 增大的程度小,优化梯度小;当 z 很大时,\hat z 增大的程度大,优化梯度大。

from dsl.

Related Issues (9)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.