Case Study

Evaluating Directions for a Startup's Anti-Spam Filter

Case context: You are leading an ML team to build a new email anti-spam system. The team proposes several distinct development directions and turns to you, expecting you to confidently select the single best direction because of your strong general ML background. However, you lack specific expertise in email filtering.

Question: Based on Andrew Ng's advice, how should you assess your ability to confidently choose the single best initial direction in this scenario?

Sample answer: You should acknowledge that it will be very difficult to confidently pick the single best initial direction. According to Andrew Ng, even someone with extensive experience in anti-spam would have a hard time picking the right direction from several ideas. Because you are not an expert in the specific application area, choosing the correct initial path will be even harder.

Key points:

  • Confidently picking the single best initial direction is highly unlikely.
  • Even experts in the application area struggle to choose among several ideas.
  • Lacking domain expertise makes picking a direction even more difficult.

Rubric: The response must correctly diagnose that picking the best direction is difficult, reference that even domain experts struggle with this, and apply the fact that lacking specific application area expertise makes it even harder.

0

1

Updated 2026-05-27

Contributors are:

Who are from:

Tags

Machine Learning

Deep Learning

Machine Learning Strategy

Supervised Learning

Dive into Deep Learning @ D2L

Data Science

Machine Learning Yearning @ DeepLearning.AI

Related