Corpus ID: 182952446

Neural Spline Flows

@article{Durkan2019NeuralSF,
  title={Neural Spline Flows},
  author={Conor Durkan and A. Bekasov and I. Murray and George Papamakarios},
  journal={ArXiv},
  year={2019},
  volume={abs/1906.04032}
}
  • Conor Durkan, A. Bekasov, +1 author George Papamakarios
  • Published 2019
  • Mathematics, Computer Science
  • ArXiv
  • A normalizing flow models a complex probability density as an invertible transformation of a simple base density. Flows based on either coupling or autoregressive transforms both offer exact density evaluation and sampling, but rely on the parameterization of an easily invertible elementwise transformation, whose choice determines the flexibility of these models. Building upon recent work, we propose a fully-differentiable module based on monotonic rational-quadratic splines, which enhances the… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    Graphical Normalizing Flows
    Invertible Generative Modeling using Linear Rational Splines
    3
    Generative Flows with Matrix Exponential
    Density Deconvolution with Normalizing Flows
    Stochastic Normalizing Flows
    3
    OT-Flow: Fast and Accurate Continuous Normalizing Flows via Optimal Transport
    1
    Equivariant Flows: exact likelihood generative learning for symmetric densities
    A RAD approach to deep mixture models
    16
    Gradient Boosted Flows

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 68 REFERENCES
    Cubic-Spline Flows
    12
    FFJORD: Free-form Continuous Dynamics for Scalable Reversible Generative Models
    211
    Improved Variational Inference with Inverse Autoregressive Flow
    737
    Deep Diffeomorphic Normalizing Flows
    11
    Sum-of-Squares Polynomial Flow
    35
    Block Neural Autoregressive Flow
    33
    Neural Autoregressive Flows
    135
    Sylvester Normalizing Flows for Variational Inference
    114
    Autoregressive Energy Machines
    12
    NICE: Non-linear Independent Components Estimation
    594