67. How would you evaluate an active-learning system that chooses the next research experiment?
HardAI-assisted research
Technical
Show model answer
Assess whether its objective matches the scientific question and how it balances informative exploration with promising candidates. Examine uncertainty quality, constraints, coverage and the possibility of repeatedly sampling a biased region.
Compare its recommendations with an appropriate baseline and review outcomes independently. Keep qualified researchers responsible for safe, authorised experiment selection; a model choosing an informative point is not evidence that the proposed activity is permissible or scientifically decisive.
What interviewers listen for
- Align acquisition with the research objective
- Check uncertainty and search coverage
- Retain qualified selection responsibility