From the course: Foundations of Responsible AI
Join today to access over 24,800 courses taught by industry experts.
Model selection trade-offs
From the course: Foundations of Responsible AI
Model selection trade-offs
- Selecting a model always comes with trade-offs. It's rarely about maximizing a single metric, but rather about matching a model to a task, to the context it'll operate in, and to the constraints that we're often working with. This video focuses on how to approach model selection when responsible AI is a part of the equation. Most teams already weigh performance and latency and cost, but when the goal is to build systems that are reliable in context, in real-world use, not just perform well in test environments, other factors become essential. They include interpretability, robustness across user groups, and the ability to maintain that model over time. Let's look at four key questions that can guide model selection, and how those questions translate into actual decisions for technical leads. The first is, what is the system's intended function, and what kind of behaviors acceptable in the real world? Imagine you're deploying a classifier to triage customer support requests for a…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.