Simulations

we first construct a simulation platform with disease evolution and treatment reactions modeled for Cancer and Sepsis, and then conduct experiments to evaluate treatment outcomes of different recommendation approaches.

Dosage Recommendation for Cancer

As shown in Fig.1, we use the mathematical model proposed by [1] to simulate the general cancer evolution and drug treatment effects.

Fig. 1. Cancer Simulator: the state transition during 6-month simulation depends on last-step feature values, current dosage action and survival analysis.

Blood Purification for Sepsis

As shown in Fig.2, We employ the mathematical model derived by [2] to simulate the acute inflammation process in response to an infection.

Fig.2. Sepsis Simulator: 8 of the 19 physiological state features are observable to the agent and the action only affects a subset of them: Na, PI, AI.

Page updated

Google Sites

Report abuse