RL in healthcare — key challenges.
hardAnswer
- (1) Sample efficiency (real trials cost billions).
- (2) Safety absolute — can't try random treatments.
- (3) Off-policy evaluation on historical patient data (dosing decisions from EHR) is standard.
- (4) Confounders in observational data — need causal RL.
- (5) Reward specification: what is 'good outcome' over years?
- Applications: dosing optimization (mechanical ventilation, sepsis), personalized treatment.
- Deployed in advisory tools only; not autonomous treatment selection.
Check yourself — multiple choice
- Same as games
- Sample eff + absolute safety + off-policy eval on historical EHR + confounders (causal RL) + long-horizon reward; deployed as advisory (ventilation, sepsis) not autonomous
- Random
- Not real
Healthcare RL: sample eff + safety + OPE + causal + long-horizon reward; advisory only.
#applications#safety
Practise Reinforcement Learning
214 interview questions in this topic.