ML笔记:(四) Theory of Generalization

Effective Number of Lines 以下討論時,ML是否可行。考慮以下條件: 即如下圖所示: 可見與圖中處時,而由藍色區域結合uniform distribution的CDF判斷 。 令 即 即只要,會大於設定的threshold。 所以: 這邊我們的有無數多個,可...

MLG笔记:(七) Node Embeddings

Traditional ML for Graphs 在傳統的Graph ML中我們通過Feature engineering 處理原始輸入的graph得到Structured Feature,再將這些結構化的特征輸入ML演算法中得到可供預測用的機器學習模型。 Graph Representat...

ML笔记:(三) Feasibility of Learning

Infeasibility of Learning All learning algorithm has its assumptions behind. No algorithm is best for all learning problems. 上圖中 對於data set中的所有個例...

ML笔记:(二) The Learning Problems

Outline Lecture 2: The Learning Problems Learning with Different Output Space Learning with Different Data Label Learning with Different Protocol ...

ML笔记:(一) Basics of Machine Learning

前言 本筆記來自台大林軒田老師的CSIE5043 Machine Learning 2021FALL課程,於此做一個學習記錄。 What is Machine Learning skill: improve some performance measure(e.g. prediction ...

MLG笔记:(六) Traditional feature-based methods: Graph-level features

Goal: We want features that characterize the structure of an entire graph Background: Kernel Methods Idea: Design kernels instead of feature vectors...

MLG笔记:(五) Traditional feature-based methods: Link-level features

Link-level Prediction: predict new links based on existing links. The key is to design features for a pair of nodes. > Two formulations of the ...

MLG笔记:(四) Traditional feature-based methods: Node-level features

Motivation We wanna be able to create additional features that will describe how this particular node is positioned in the rest of the network, and wha...

MLG笔记:(三) Choice of Graph Representation

Components of a Network(graph) Object: nodes,vertices   Interactions: links,edges   System: network,graph   Types of graph 有向圖與無向圖的選擇: ...

MLG笔记:(二) Applications of Graph ML

Different tasks of GraphML Node classification: Predict a property of a node.(e.g. Categorize online users) Link prediction: Predict whether there are...