Machine Learning and the Professions

Royal Society, London

Neil D. Lawrence

University of Sheffield

@lawrennd inverseprobability.com

What is Machine Learning?

  • A set of techniques for extracting information from data.

  • Principal technology underlying recent advances in artificial intelligence.

  • A major technology underlying data science.

Example: Image Classification

Perceptron

Example: AlphaGo

  • An achievement in Artificial Intelligence

Example: Driverless Cars

  • Engineered system with many components.
  • Many vital components are based on machine learning.
  • Pedestrian recognition (an image classification problem)

Example: Facebook

  • Most aspects of the interface controlled using machine learning.

  • Ranking of posts in newsfeed.

  • Ranking of adverts.

Machine Learning

data + model = algorithm
  • Your actions (likes, friends, clicks) are turned into a vector.

\[\mathbf{x}=\begin{bmatrix} 1.0 & 2.3 & 3 \end{bmatrix}\]

You Will Need

  • Training data

  • A model of how they interact

  • Some computer time

The Professions

  • What aspects cannot/can be replaced by a machine?

    • What makes professions special.

    • Encroachment of automation on ‘middle class’ roles.

  • Not just your role, but the wider societal role.

  • How will the professions evolve?

Thanks!