Reinforcement Learning for Respondent-Driven Sampling
Authors:
Justin Weltz,
Angela Yoon,
Yichi Zhang,
Alexander Volfovsky,
Eric Laber
Abstract:
Respondent-driven sampling (RDS) is widely used to study hidden or hard-to-reach populations by incentivizing study participants to recruit their social connections. The success and efficiency of RDS can depend critically on the nature of the incentives, including their number, value, call to action, etc. Standard RDS uses an incentive structure that is set a priori and held fixed throughout the s…
▽ More
Respondent-driven sampling (RDS) is widely used to study hidden or hard-to-reach populations by incentivizing study participants to recruit their social connections. The success and efficiency of RDS can depend critically on the nature of the incentives, including their number, value, call to action, etc. Standard RDS uses an incentive structure that is set a priori and held fixed throughout the study. Thus, it does not make use of accumulating information on which incentives are effective and for whom. We propose a reinforcement learning (RL) based adaptive RDS study design in which the incentives are tailored over time to maximize cumulative utility during the study. We show that these designs are more efficient, cost-effective, and can generate new insights into the social structure of hidden populations. In addition, we develop methods for valid post-study inference which are non-trivial due to the adaptive sampling induced by RL as well as the complex dependencies among subjects due to latent (unobserved) social network structure. We provide asymptotic regret bounds and illustrate its finite sample behavior through a suite of simulation experiments.
△ Less
Submitted 2 January, 2025;
originally announced January 2025.
The role of viral infectivity in oncolytic virotherapy outcomes: A mathematical study
Authors:
Pantea Pooladvand,
Chae-Ok Yun,
A-Rum Yoon,
Peter S. Kim,
Federico Frascoli
Abstract:
A model capturing the dynamics between virus and tumour cells in the context of oncolytic virotherapy is presented and analysed. The ability of the virus to be internalised by uninfected cells is described by an infectivity parameter, which is inferred from available experimental data. The parameter is also able to describe the effects of changes in the tumour environment that affect viral uptake…
▽ More
A model capturing the dynamics between virus and tumour cells in the context of oncolytic virotherapy is presented and analysed. The ability of the virus to be internalised by uninfected cells is described by an infectivity parameter, which is inferred from available experimental data. The parameter is also able to describe the effects of changes in the tumour environment that affect viral uptake from tumour cells. Results show that when a virus is inoculated inside a growing tumour, strategies for enhancing infectivity do not lead to a complete eradication of the tumour. Within typical times of experiments and treatments, we observe the onset of oscillations, which always prevent a full destruction of the tumour mass. These findings are in good agreement with available laboratory results. Further analysis shows why a fully successful therapy cannot exist for the proposed model and that care must be taken when designing and engineering viral vectors with enhanced features. In particular, bifurcation analysis reveals that creating longer lasting virus particles or using strategies for reducing infected cell lifespan can cause unexpected and unwanted surges in the overall tumour load over time. Our findings suggest that virotherapy alone seems unlikely to be effective in clinical settings unless adjuvant strategies are included.
△ Less
Submitted 7 October, 2021;
originally announced October 2021.
