liveBook · Manning

Risk landscape

As explained in Machine Learning System Design

A structured view of potential failure outcomes, their consequences, and their relevant likelihoods for a system.

A risk landscape maps possible system errors to their consequences and likelihoods. It distinguishes harmless mistakes from failures that damage a business or threaten lives, making risk central to system design. section 2.4

Loan approval

A lender must compare approving a loan for a likely defaulter with rejecting a customer who would repay successfully. The losses may differ by number of people, amount lent, or changing repayment ratios. Making those tradeoffs explicit helps set reliability requirements and choose useful metrics. section 2.4

Use the risk landscape to:

  • Identify which errors the system must avoid. section 2.4
  • Translate mistake costs into reliability, data-gathering, and metric requirements. section 2.4

How Machine Learning System Design covers it

How deep it goes
Explains consequence and likelihood tradeoffs, then stops short of prescribing a specific reliability target or metric.
Examples
Contrasts approving a likely defaulting borrower with rejecting a borrower who would repay successfully.
What you can do after
Evaluate system errors and use their costs to guide reliability, data, metric, and design decisions.
How it connects
Connects risk scores and thresholds with reliability requirements, data gathering, metrics, and business costs.
Cover of Machine Learning System Design
liveBookmanning.com
Read Machine Learning System Design on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Machine Learning System Design. The section links point to the parts of the book it draws on.