The challenges of applied machine learning

Applying machine learning to real-world settings poses challenges that are not present in academic settings. You will be creating applications that will impact the work and lives of thousands and millions of people.

You’ll be dealing with technical and non-technical challenges. You’ll need a lot of data and the technical plumbing that can support your machine learning models during training and maintenance. You’ll need an IT infrastructure that can support your models as they scale.

And you’ll need business people who can define the problem and objectives. You’ll need researchers who can get input and feedback from users. And you’ll need a legal team who can spot and deal with sensitive and ethical issues.

Real World AI: A Practical Guide for Responsible Machine learning, a book by Alyssa Simpson Rochwerger and Wilson Pang, discusses the challenges of applied machine learning and provides guidelines and tips for avoiding fatal mistakes in designing and implementing machine learning strategies at your organization.

In my latest column in TechTalks, I explore four challenges discussed in Real World AI along with my own input and experience on the topic.

Read the full article here.

For more on the business of artificial intelligence:

• Why most machine learning strategies fail

• The common traits of successful AI startups

• Understanding the AI factory

• Is your business ready for artificial intelligence?