Building Guardrails Against Overconfidence [LISTEN or WATCH]

Add bookmark

Building Guardrails Against Overconfidence [LISTEN or WATCH]

Add bookmark

Data Scientist Jett Oristaglio joins us to talk about Humble AI. Trusted AI and asks, What is everything you need to trust in AI for your life?

He goes on to share that trust is not binary- it’s multi -dimensional - and so, there are a number of criteria that should be evaluated.

Performance- or, does the model predict the problem being solving for well- otherwise known as accuracy. Operations- or, how reliable is the system that the model is deployed in. And Ethics- or, how well does the model align with the values of the organization.

Of course, most talked about is accuracy in that the model has to be fast enough, has to be stable (have strength). Not surprisingly- all of these aspects, you want in a human decision maker.

Listen

Or Watch


Sponsored By:

More From DataRobot

Visual AI: See the Big Picture with Automated Computer Vision

The current goal for leading global corporate enterprises is the democratization and acceleration...

2021-02-23
Visual AI: See the Big Picture with Automated Computer Vision

Engaging Analytics & Humility to Mitigate Uncertainty

In reality, most models will be tasked with scoring unusual, anomalous, or unexpected data at some p...

2020-11-02 by Seth Adler
Engaging Analytics & Humility to Mitigate Uncertainty

Humility in AI: Achieving Optimal Performance, Accuracy & Decision-Making

With the rising stakes, AI systems must be built to be humble, just like humans. AI should know when...

2020-11-02 by Seth Adler
Humility in AI: Achieving Optimal Performance, Accuracy & Decision-Making

Engaging Analytics & Humility to Mitigate Uncertainty

In reality, most models will be tasked with scoring unusual, anomalous, or unexpected data at some p...

2020-10-21 by Seth Adler
Engaging Analytics & Humility to Mitigate Uncertainty

Sponsor Page

Recommended