Corpus ID: 199577465

Thompson Sampling and Approximate Inference

@inproceedings{Phan2019ThompsonSA,
  title={Thompson Sampling and Approximate Inference},
  author={My V T Phan and Yasin Abbasi-Yadkori and Justin Domke},
  booktitle={NeurIPS},
  year={2019}
}
  • My V T Phan, Yasin Abbasi-Yadkori, Justin Domke
  • Published in NeurIPS 2019
  • Mathematics, Computer Science
  • We study the effects of approximate inference on the performance of Thompson sampling in the $k$-armed bandit problems. Thompson sampling is a successful algorithm for online decision-making but requires posterior inference, which often must be approximated in practice. We show that even small constant inference error (in $\alpha$-divergence) can lead to poor performance (linear regret) due to under-exploration (for $\alpha 0$) by the approximation. While for $\alpha > 0$ this is unavoidable… CONTINUE READING

    Figures and Topics from this paper.

    Citations

    Publications citing this paper.
    SHOWING 1-5 OF 5 CITATIONS

    On Thompson Sampling with Langevin Algorithms

    VIEW 2 EXCERPTS
    CITES BACKGROUND

    On Thompson Sampling for Smoother-than-Lipschitz Bandits

    VIEW 1 EXCERPT
    CITES BACKGROUND

    Scalable Thompson Sampling using Sparse Gaussian Process Models

    VIEW 1 EXCERPT
    CITES BACKGROUND

    TS-UCB: Improving on Thompson Sampling With Little to No Additional Computation

    VIEW 1 EXCERPT
    CITES BACKGROUND

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 18 REFERENCES

    Further Optimal Regret Bounds for Thompson Sampling

    VIEW 2 EXCERPTS

    Ensemble Sampling

    Thompson Sampling for Complex Bandit Problems

    VIEW 5 EXCERPTS
    HIGHLY INFLUENTIAL

    Variational Inference: A Review for Statisticians

    VIEW 4 EXCERPTS
    HIGHLY INFLUENTIAL

    An Information-Theoretic Analysis of Thompson Sampling

    VIEW 3 EXCERPTS

    Improving the Expected Improvement Algorithm

    VIEW 7 EXCERPTS
    HIGHLY INFLUENTIAL

    Simple Bayesian Algorithms for Best Arm Identification

    VIEW 5 EXCERPTS
    HIGHLY INFLUENTIAL