The Discipline of Machine Learning 笔记

Tom M. Mitchell于2006年所写对Machine Learning进行简单介绍的论文

Defining Questions

Machine Learning的核心问题是:How can we build computer systems that automatically improve with experience, and what are the fundamental laws that govern all learning processes?

这里的learn的定义是:with respect to a particular task T, performance metric P, and type of experience E, if the system reliably improves its performance P at Task T, following experience E.

Machine Learning是Computer Science和Statistics的交集,还有一个较弱相关领域就是the study of human and animal learning in Psychology, Neuroscience and related fields,其他对Machine Learning感兴趣的领域如biology、ecomonics和control theory。

State of Machine Learning

Application Successes

  • Speech recognition.
  • Computer vision.
  • Bio-surveillance. 比如detect and track disease outbreak。
  • Robot control. 比如无人飞行器。
  • Accelerating empirical sciences.

Place of Machine Learning within Computer Science

在如下领域Machine Learning已经是best practice。即complex和self-customizing的软件。

  • The application is too complex for people to manually design the algorithm. 比如sensor-base perception tasks (speech recognition, computer vision, etc.)
  • The application requires that the software customize to its operational environment after it is fielded. 比如speech recognition应该用用户数据再训练一遍,推荐系统,自定义垃圾邮件。

Some Current Research Questions

正在被研究(2006年)的前沿问题,例如:

  • Can unlabeled data be helpful for supervised learning?
  • How can we transfer what is learned for one task to improve learning in other related tasks?
  • What is the relationship between different learning algorithms, and which should be used when?
  • For learners that actively collect their own training data, what is the best strategy?
  • To what degree can we have both data privacy and the benefits of data mining?

Longer Term Research Questions

  • Can we build never-ending learners?
    按照进阶顺序不断学习各种课程,课程之间可以融会贯通。 A key research issue here is self-supervised learning and constructing an appropriate graded curriculum.
  • Can machine learning theories and algorithms help explain human learning?
  • Can we design programming languages containing machine learning primitives?
  • Will computer perception merge with machine learning?

Ethical Questions

personal privacy.

Where to Learn More

The top conferences and journals in the field:

  • International Conference on Machine Learning (ICML).
  • Conference on Neural Information Processing Systems (NIPS).
  • Annual Conference on Learning Theory (COLT).
  • Journal of Machine Learning Research (JMLR). This top journal is freely available online at www.jmlr.org.
  • Machine Learning. Published by Springer.