gRc402 • Personalization
& Mass-Customization

Week 10 • Winter 2021 • Ahmed Sagarwala

Grading submission

  • Not entertaining resubmissions
  • About 20-30 hours of grading to go
  • 'Final grades' on Canvas

Final Project Submission (March 19)

  • Submit by March 15 for quick turnaround
  • 15% of your final grade
  • Campaign overview (workflow, KPIs, team, etc.)
  • Postcard w/ 3+ samples
  • Anything else from workflow w/ 3+ samples
  • Show personalization and branding (annotated image)
  • See Canvas for guidelines

Objectives

  • Provide an overview of algorithms
  • Develop a working model of an algorithm
  • Learn how machine learning is implemented
  • Determine why personalization fails when using AI
  • Discuss why AI cannot easily be audited

What's your approach?

Write your down steps to solving a puzzle (5 min)
Discuss steps as a group (5 min)

 

Machine Learning

  • Subset of artificial intellegence
  • Requires a dataset*
  • Pattern recognition
  • Supervised vs unsupervised learning
  • Neural networks

Input

Output

Probability

Hidden layer

x1 = Height

x2 = Weight

o1 = Gender

Image: Victor Zhou

Deep fakes & Police States

  • It's getting harder to believe what we see
  • If you look for it, you'll find it
  • Confirmation bias
  • Sensors and hand-washing
  • Webcams and facial recognition
  • "WMDs"
  • Never leave a computer to do pass judgement
  • Consider your implicit bias in the design

Use data for good

Music recommendations

We want to create an algorithm that will provide recommendations to a listener.

 

What could be our inputs?

What could be our outputs?

GrC402 Week 10

By Am Sagarwala

GrC402 Week 10

Alogrithms, AI, and personalization

  • 327

More from Am Sagarwala