: Alyssa Simpson Rochwerger, Wilson Pang
: Real World AI A Practical Guide for Responsible Machine Learning
: Lioncrest Publishing
: 9781544518824
: 1
: CHF 10.50
:
: Sonstiges
: English
: 220
: kein Kopierschutz
: PC/MAC/eReader/Tablet
: ePUB
How can you successfully deploy AI? When AI works, it's nothing short of brilliant, helping companies make or save tremendous amounts of money while delighting customers on an unprecedented scale. When it fails, the results can be devastating. Most AI models never make it out of testing, but those failures aren't random. This practical guide to deploying AI lays out a human-first, responsible approach that has seen more than three times the success rate when compared to the industry average. In Real World AI, Alyssa Simpson Rochwerger and Wilson Pang share dozens of AI stories from startups and global enterprises alike featuring personal experiences from people who have worked on global AI deployments that impact billions of people every day. AI for business doesn't have to be overwhelming. Real World AI uses plain language to walk you through an AI approach that you can feel confident about-for your business and for your customers.

Introduction


Alyssa


In late 2015, as a product manager within the newly formed computer vision team at IBM®, we were days away from launching the team’s firstcustomer-facing output. For months, we’d been working to create a commercially availablevisual-recognition application programming interface (API) that more than doubled the accuracy of existing models. The company had high hopes for scaling the API into a significant revenue stream. Our biggest focus to date had been improving the model’s F1 score—a standard academic measure of a classification system’s accuracy—against a subset of our training data, which included tens of millions of images and labels the team had compiled over months and years.

The API was meant to be used to tag images fed into it with descriptive labels. For example, you could feed it an image of a brown cat, and it would return a set of tags that would include “cat,” “brown,” and “animal.” Businesses would be able to use it for all kinds of applications—everything from building user preference profiles by scraping images posted to social media, to ad targeting, or customer experience improvements. Over the past several months, to train and test the system, we’d used over 100 million images and labels from a variety of sources as training data. We’d succeeded in improving the F1 score considerably, to the point where an image I fed it of my sister and me at a wedding immediately came back taggedbridesmaids, which I thought was impressive.

And now, with all of IBM’s release checklists completed and a planned launch mere days away, I was faced with an unanticipated problem.

That morning, I received a message from one of our researchers that washeart-stopping in its simple urgency:We can’t launch this. When I asked why, he sent me a picture of a person in a wheelchair that he’d fed into the system as a test. The tag that came back?

Loser.

Panic. IBM has a 100-year history of inclusion and diversity. So, besides being objectively horrible, this output clearly indicated that the system did not reflect IBM’s values. While we had beenlaser-focused on improving the system’s accuracy, what other types of harmful and unintended bias had we accidentally introduced?

I immediately sounded the alarm, alerted my bosses, and scrubbed the launch. Our team got to work. Besides fixing the model, we had two main questions