Delegative system

As explained in Delivering Machine Learning Projects: From design to deployment

A delegative system is a system that makes decisions on a user's behalf while providing the human with timely mechanisms to review, correct, or override those decisions.

A delegative system makes decisions for users when case-by-case human control is impractical because decisions require overwhelming data or rapid response. It preserves human review, correction, and intervention when automated behavior fails. section 9.2.2

Choose delegation deliberately because it changes functional and nonfunctional requirements. Production delivery must provide suitable data flows, model serving, user controls, instrumentation, logging, monitoring, alerting, testing, sign-off, and ethical review. section 9.11

Perseverance rover

The rover independently proposes a terrain path, sends it to Mission Control for approval, prepares the command process, and executes the approved plan. This balances communication latency against the risk of full autonomy: manual control would be too slow, while complete autonomy would expose the mission to unacceptable risk. section 9.2.2

Figure 9.3 The control flow for a delegative remote control rover [1]

Use the control-flow diagram to trace how independent planning is separated from human approval and execution. section 9.2.2

How Delivering Machine Learning Projects: From design to deployment covers it

How deep it goes

Defines the pattern, works through episodic rover control, and connects the choice to production requirements without detailing implementation algorithms.

Examples

Uses the Perseverance rover to demonstrate proposed paths, Mission Control approval, command preparation, and independent execution.

What it assumes

Assumes familiarity with ML system delivery and distinguishes delegation from assistive and autonomous operation.

What you can do after

Operate/configure a delegated system by specifying user controls, approval points, observability, testing, and production sign-off.

How it connects

Connects human oversight, episodic control, autonomy, risk management, model serving, monitoring, and ethical review.

liveBookmanning.com

Read Delivering Machine Learning Projects: From design to deployment on liveBook Get the book at manning.com

Related concepts

• episodic controlused by · Episodic control is an autonomy-management technique in which a system independently performs bounded episodes of activity subject to human approval or intervention between episodes.
• autonomous systemcontrasts with · An autonomous system is a collection of networks and routers under a common administrative control that participates in interdomain routing.
• partitioning strategyalternative to · A partitioning strategy divides a system into independent subsystems so that an individual subsystem can fail without causing systemic or unacceptable damage.
• machine learning systemkind of · A machine learning system is a software system that uses models trained on data, together with associated data, computation, and operational components, to perform tasks.
• fake delegative systemincludes · A fake delegative system is a system that nominally transfers control to a human but does so too late or with insufficient authority to prevent harm.

This explainer was generated by Manning's knowledge graph from the text of Delivering Machine Learning Projects: From design to deployment. The section links point to the parts of the book it draws on.