Ikibazo gihuza ibibazo · Hagati Urwego

Ikizamini cy'Ibisobanuro by'Impamo ku Byemezo bya AI

Tandukanya icyitegererezo gishingiye ku ngaruka zishingiye ku mpamvu zishingiye ku mpamvu no gusuzuma niba bishoboka, ibikorwa, inzitizi, n'ibikenewe mu bujurire.

Inzira zijyanye no kuyoboraIbisobanuro Binyuranye Kubyemezo bya AI
Ikibazo 1 Bya 8

What question does a counterfactual explanation answer?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. What question does a counterfactual explanation answer?
    • What small input change would make the model produce a different output?
    • What source code created the model?
    • What decision will a court make?
    • What data are stored on a server?
  2. Does a counterfactual prove that changing a feature will cause a real-world outcome?
    • Yego, burigihe
    • No; it describes model behavior, not necessarily real-world causation
    • Only for credit decisions
    • Only if the model is interpretable
  3. A counterfactual says a person would qualify if they were ten years younger. What concern does that raise?
    • It is necessarily a helpful action
    • Age is always easy to change
    • The change may be immutable, infeasible, or legally problematic
    • It reveals the full model source code
  4. Why constrain counterfactual feature combinations?
    • To avoid examples that are implausible or impossible
    • To make all explanations identical
    • To prevent model validation
    • To increase the number of protected features
  5. How should a counterfactual about a loan balance be presented?
    • As a guaranteed loan approval
    • As proof of the lender’s causal reasoning
    • As a model-based contrast that is not a promise of approval
    • As a replacement for all adverse-action duties
  6. Which quality matters when a counterfactual is meant to help a person take action?
    • Feasibility and actionability
    • A large number of technical terms
    • Icyitegererezo cyamahugurwa igihe
    • Whether it hides errors
  7. What should happen if a counterfactual recommends changing a protected or immutable trait?
    • Treat it as ordinary advice
    • Ignore it without review
    • Escalate it as a possible fairness or legal design issue
    • Tell the person the change is mandatory
  8. Which process complements counterfactual explanations in consequential decisions?
    • Data correction, human review, and appeal routes
    • Removing all decision records
    • Relying only on the generated explanation
    • Making the model score confidential