Leture 1: The Learning Problem

沥川 bio photo By 沥川

The learning problem - Outline

  • Example of machine learning
  • Components of Learning
  • A simple model
  • Types of learning
  • Puzzle

Example: Predicting how a viewer will rate a movie

  • 10% improvement = 1 million dollar prize, 这是机器学习的商业价值, 一点小小的调优, 可以更合理的分配资源, 提升效率, 得到巨大经济利益.
  • The essence of machine learning:
    • A pattern exists. 机器学习如果要成立, 模式/规律必须存在, 完全随机无学习可言
    • We cannot pin it down mathematically. 人类无法找到数学规律, 需要借助计算机强大的计算能力, 和模式识别能力, 来帮助人类找到模型.
    • We have data on it. 机器学习其实是从数据中学习, 或者转化成数字, 进行处理.

Movie rating - a solution

  • viewer: like movies? / like action? / prefers blockbusters?… likes Tom Cruise?
  • movie: comedy content / action content / blockbusters? … Tom Cruise in it?
  • 结合 viewer 和 movie 的特征, 建立学习算法, 得到评价分数

Components of learning

  • Formalization:
    • Input: X (customer application)
    • Output: y (good/bad customer?)
    • Target function: f: X -> y (ideal credit approval formula)
    • Data: $(x_1, y_1), (x_2, y_2), … , (x_n, y_n)$ (historical records)
    • Hypothesis: g: X -> y (formula to be used)
  • Solution components
  • The 2 solution components of the learning problem:
    • The Hypothesis Set
    • The Learning Algorithm
  • Together, they are referred to as the learning model.

A simple hypothesis set - the ‘perceptron’

  • For input $X = (x_1, … , x_d)$ ‘attributes of a customer’
    • Approve credit if $\sum_{i=1}^{d}w_ix_i$ > threshold,
    • Deny credit if $\sum_{i=1}^{d}w_ix_i$ < threshold.
  • This linear formula $h \in H$ can be written as
    • h(x) = sign(($\sum_{i=1}^{d}w_ix_i$) - threshold)

Types of learning

Basic premise of learning

“using a set of observations to uncover an underlying process”

broad premise => many variations

  • Supervised Learning 有监督学习, 有先验知识
  • Unsupervised Learning 无监督学习, 聚类..
  • Reinforcement Learning 增强学习, 玩游戏, 延迟反馈

Changelog

  • 2018-02-26 创建