Diffusion Approximations for Thompson Sampling in the Small Gap Regime
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2021
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911613062217728 |
|---|---|
| author | Fan, Lin Glynn, Peter W. |
| author_facet | Fan, Lin Glynn, Peter W. |
| contents | We study the process-level dynamics of Thompson sampling and related sampling-based bandit algorithms in the ``small gap'' regime, where the gaps between the arm means are of order $\sqrtγ$ or smaller and the time horizon is of order $1/γ$, with $γ\downarrow 0$. In this regime, as $γ\downarrow 0$, we show that the process-level dynamics of such algorithms converge weakly to the solutions to certain stochastic differential equations and stochastic ordinary differential equations. Our weak convergence theory is developed using the Continuous Mapping Theorem, which provides a direct and modular theoretical approach that can be adapted to analyze a variety of sampling-based bandit algorithms and handle weakly dependent reward processes. A central finding is an algorithmic invariance principle: in the small gap regime, the limit dynamics of a broad class of sampling-based algorithms -- including Thompson sampling with general single-parameter exponential family likelihoods, as well as non-parametric bandit algorithms based on bootstrap re-sampling -- all coincide with those of Thompson sampling with Gaussian likelihoods. Moreover, in the small gap regime, the regret performance of these algorithms is generally insensitive to model mis-specification, changing continuously with increasing degrees of mis-specification. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2105_09232 |
| institution | arXiv |
| publishDate | 2021 |
| record_format | arxiv |
| spellingShingle | Diffusion Approximations for Thompson Sampling in the Small Gap Regime Fan, Lin Glynn, Peter W. Machine Learning Statistics Theory 62B15, 60J70 We study the process-level dynamics of Thompson sampling and related sampling-based bandit algorithms in the ``small gap'' regime, where the gaps between the arm means are of order $\sqrtγ$ or smaller and the time horizon is of order $1/γ$, with $γ\downarrow 0$. In this regime, as $γ\downarrow 0$, we show that the process-level dynamics of such algorithms converge weakly to the solutions to certain stochastic differential equations and stochastic ordinary differential equations. Our weak convergence theory is developed using the Continuous Mapping Theorem, which provides a direct and modular theoretical approach that can be adapted to analyze a variety of sampling-based bandit algorithms and handle weakly dependent reward processes. A central finding is an algorithmic invariance principle: in the small gap regime, the limit dynamics of a broad class of sampling-based algorithms -- including Thompson sampling with general single-parameter exponential family likelihoods, as well as non-parametric bandit algorithms based on bootstrap re-sampling -- all coincide with those of Thompson sampling with Gaussian likelihoods. Moreover, in the small gap regime, the regret performance of these algorithms is generally insensitive to model mis-specification, changing continuously with increasing degrees of mis-specification. |
| title | Diffusion Approximations for Thompson Sampling in the Small Gap Regime |
| topic | Machine Learning Statistics Theory 62B15, 60J70 |
| url | https://arxiv.org/abs/2105.09232 |