betaFeedback

WorksComputer Science

Two-Phase Rapid Simulation-Based Inference with Differentiable Simulators

Abstract

We introduce Rapid Simulation Based Inference (RSBI), a diffusion-based variational approach to likelihood-free Bayesian inference that achieves high sampling efficiency under large prior-to-posterior volumes. RSBI builds on advances in Schrödinger Bridge (SB) diffusion sampling to handle multi-modal posteriors, with an initial likelihood supplied by a distance measure via a differentiable simulator. The key innovation of this initial surrogate likelihood is that it gives a highly efficient proposal distribution, replacing sequential rounds of other methods with an efficient one-shot proposal. Subsequent refinement then transports the proposal to the true posterior as specified by the simulator and the original prior. The SB sampler is warm-started from the proposal, and its energy is corrected by a neural likelihood ratio that is learned only where the proposal places mass. As a variational method, RSBI does not suffer from prior leakage commonly observed in standard sequential posterior estimation methods. To improve mode coverage we optionally utilize the well-tempered meta-dynamics framework, which also encourages exploration of the prior volume. We observe strong performance on standard SBI benchmarks, particularly when the prior volume is increased relative to the ground truth posterior, achieving a performance degradation of only ∼5%\sim5\% on Two Moons under a highly constrained simulation budget. Additionally, we evaluate RSBI's performance for gravitational wave ring-down posterior estimation, in a real-world physics benchmark inspired by black hole spectroscopy.

Introduction

Many domains in science and engineering rely on stochastic simulators whose sampling distribution p(x∣θ)p(x \mid\theta) is intractable but from which data samples xx can be drawn given parameters θ\theta. Bayesian parameter inference in such settings is the subject of Simulation-Based Inference (SBI) [1]. Neural SBI methods have matured rapidly over the past five years. Sequential Neural Posterior Estimation [SNPE; [2, 3]] targets the posterior, Sequential Neural Likelihood Estimation [SNLE; [4]] the likelihood, and Sequential Neural Ratio Estimation [SNRE; [5, 6]] the likelihood-to-evidence ratio. Sequential Neural Variational Inference [SNVI; [7]] combines these with variational inference (VI).

A common difficulty arises when the prior-to-posterior volume ratio (Rp\mathcal{R}_{p}) is large: when p(θ)p(\theta) is broad compared to the posterior, most simulator calls fall in regions of negligible posterior density, and efficiency vanishes. Sequential variants can reduce this cost by retraining the proposal across rounds, but this comes at a price. SNPE with atomic proposals suffers from posterior leakage outside prior support [3], and naive implementations of variational approaches collapse onto a single mode of the posterior [7]. Truncation schemes [8, 9, 10] instead restrict the effective prior, so their result is exact only if the truncated region contains the posterior, or else must be corrected at inference time, for example by importance sampling [11].

A growing class of scientific simulators are differentiable: GPU-native models written within the JAX [12]/PyTorch [13] frameworks, learnable physical simulators, and neural emulators trained to mimic expensive codes all expose the gradient ∇θs(θ)\nabla_{\theta} s(\theta) of the parameters θ\theta under the simulator output, s(θ)s(\theta). This represents an opportunity for SBI: algorithms can exploit gradients directly rather than treating the simulator as a black box.

Recent advances in neural samplers, in particular the Adjoint Schrödinger Bridge Sampler (ASBS; [14]), provide an opportunity to exploit such gradients. ASBS recasts sampling from an unnormalised target distribution as a Schrödinger bridge problem solved via stochastic optimal control: a controlled SDE transports a tractable source distribution to the target, with the control trained by adjoint matching and a learned corrector [15]. The recent Well-Tempered variant (WT-ASBS; [16]) augments ASBS with a metadynamics-style bias [17], overcoming free-energy barriers between modes during training and enabling broad mode coverage at fixed simulation cost. Two properties make this family particularly well-suited to SBI. First, ASBS queries the target only through gradients of its energy (the negative log of the unnormalised target density), ∇θE\nabla_{\theta} E, which a differentiable simulator supplies. Second, the well-tempered bias addresses the mode-seeking pathology that limits standard variational SBI [7] on multi-modal posteriors. To our knowledge, neither ASBS nor its well-tempered extension have previously been deployed for likelihood-free Bayesian inference.

We introduce RSBI, a two-phase SBI framework that leverages differentiable simulators while keeping the target posterior fixed as p(θ∣xo)p(\theta\mid x_o), where xox_o denotes the observed data. Phase 1 places posterior mass rapidly using a cheap surrogate likelihood, parameterized by bandwidth ϵ\epsilon and built from distance metric Dϵ(s,xo)\mathcal{D}_{\epsilon}(s,x_o) between observed and simulated data, plugged into a Well-Tempered Adjoint Schrödinger Bridge Sampler [14, 16] (WT-ASBS). Phase 2 refines the resulting Phase-1 proposal using WT-ASBS, augmented by an NRE trained on simulator samples drawn from the ‘zoomed in’ proposal. Figure 1 outlines the process.

Actual diffusion trajectories from the RSBI pipeline on the Two Moons task

Figure 1. Actual diffusion trajectories from the RSBI pipeline on the Two Moons task detailed in Sec.5. RSBI is able to produce a high quality posterior despite the large prior size.

Our contributions are threefold:

  • A two-phase SBI method for differentiable simulators. We introduce RSBI (Sec. 4), a diffusion-based variational approach to likelihood-free Bayesian inference that achieves high sampling efficiency under large prior-to-posterior volume ratios.

  • First application of WT-ASBS to SBI. We adapt the well-tempered metadynamics extension of ASBS [16] to likelihood-free Bayesian inference, countering the mode-seeking pathology of reverse-KL variational SBI [7] on multimodal posteriors.

  • Benchmarking robustness at large prior-to-posterior volume ratio Rp\mathcal{R}_p. We benchmark RSBI and other methods on Two Moons [18], SLCP [18, 19], Lotka–Volterra [18], Bernoulli GLM [18], and a new differentiable gravitational-wave ringdown task (dim θ=6\theta= 6, dim x=820x = 820), sweeping prior-to-posterior volume ratios up to log⁡10Rp≈12\log_{10} \mathcal{R}_p \approx12.

Related Work

RSBI sits at the intersection of three research threads. The first is neural simulation-based inference, in which a density or ratio estimator is trained on simulated (θ,x)(\theta, x) pairs to target the posterior, likelihood, or likelihood-to-evidence ratio [2, 3, 4, 5, 6, 20]. Sequential variants iteratively refine their proposal to concentrate simulations in informative regions of parameter space. SNVI+fKL [7] pairs likelihood(-ratio) estimation with VI and is our closest methodological sibling. Like RSBI, SNVI combines a learned neural likelihood ratio estimated via Neural Ratio Estimation (NRE) with a variational sampler, and, like RSBI, it explicitly addresses the mode-seeking pathology of standard reverse-KL VI, in its case by training under the mass-covering forward KL with sampling–importance-resampling for proposal updates. RSBI instead handles mode coverage through the well-tempered metadynamics bias of WT-ASBS [16], which flattens free-energy barriers between modes. The two methods also differ in how the initial proposal is obtained: SNVI draws its first round of simulations from the original prior, whereas RSBI is initialised by a one-shot surrogate-guided Schrödinger bridge.

The second thread is the surrogate-likelihood lineage that connects classical ABC [21] to modern VI. Any differentiable distance metric Dϵ(s,xo)D_{\epsilon}(s, x_o) induces a pseudo-likelihood whose gradients propagate through a simulator, an idea explored by ABC-VI [22], as well as score-based approaches [23, 24]. RSBI borrows from this for Phase 1, but rather than amortising over xx or running an MCMC against the surrogate, we use it to construct the energy on which WT-ASBS is trained. This gives a non-mode-seeking proposal that Phase 2 then transports to the true posterior, with a likelihood specified by an NRE learned only where the proposal places mass. This use of a surrogate kernel as a warm-start for a downstream refinement, rather than as the inference target itself, can alleviate the bandwidth-versus-bias trade-off that limits standalone ABC-VI.

The third thread is the rapidly maturing ecosystem of GPU-native, auto-differentiable simulators across many scientific fields and neural emulators [25, 26, 27, 28]. Direct access to simulation gradients ∇θs(θ)\nabla_{\theta}s(\theta) enables gradient-based inference for models that previously admitted only analytical-likelihood treatments: ASBS [14] queries its target exclusively through energy gradients, which the differentiable simulator supplies via a distance metric Dϵ(s,xo)D_{\epsilon}(s, x_o), sidestepping the density-estimation step of standard SNP/ESNLE/SNRE. A growing number of works that exploit gradients have been used to improve neural SBI. Brehmer et al. Brehmer et al. [29] augment the training losses of ratio and likelihood estimators with the joint score and likelihood ratio, and relaxations such as the Gumbel-Softmax [30] extend such gradients to simulators with discrete randomness. To our knowledge, the only gradient-informed NPE method adds a score-matching term to the NPE loss [31], leaving the proposal, the simulation allocation and the sampler unchanged. In all of these, gradients regularise an estimator fit to prior-wide simulations. In RSBI they instead drive the Phase 1 sampler directly, through an auto-encoder-like comparison of simulations with the observation.

Background

Simulation Based Inference

We consider a stochastic simulator s:Θ×Ω→Xs : \Theta\times\Omega\to\mathcal{X} with parameters θ∈Θ⊆Rd\theta\in\Theta\subseteq\mathbb{R}^{d} and outputs x=s(θ,ω)∈Xx = s(\theta,\omega) \in\mathcal{X}, where ω∈Ω\omega\in\Omega collects all internal randomness of the simulator (noise draws, random seeds) and is drawn from a distribution p(ω)p(\omega) that does not depend on θ\theta. The simulator implicitly defines the conditional density p(x∣θ)p(x \mid\theta) and the distribution of s(θ,ω)s(\theta,\omega) under ω∼p(ω)\omega\sim p(\omega). The likelihood θ↦p(xo∣θ)\theta\mapsto p(x_o \mid\theta) is this density evaluated at the observed data xox_o. It can be sampled by running the simulator but not evaluated. We call the simulator differentiable if θ↦s(θ,ω)\theta\mapsto s(\theta,\omega) is differentiable for every ω\omega, so that pathwise gradients ∇θs(θ,ω)\nabla_{\theta}s(\theta,\omega) are available. We write s(θ)s(\theta) for s(θ,ω)s(\theta,\omega) with ω∼p(ω)\omega\sim p(\omega), and s(θ,ω1:K)={s(θ,ωk)}k=1Ks(\theta,\omega_{1:K}) = \{s(\theta,\omega_k)\}_{k=1}^{K} with ωk∼iidp(ω)\omega_k \stackrel{\mathrm{iid}}{\sim} p(\omega) for KK independent replicates at the same θ\theta. Given a prior p(θ)p(\theta) and observation xox_o, simulation-based inference (SBI) targets the posterior p(θ∣xo)∝p(xo∣θ)p(θ)p(\theta\mid x_o) \propto p(x_o \mid\theta)p(\theta) using only forward simulations [1].

Classical approximate Bayesian computation (ABC) accepts prior draws θ∼p(θ)\theta\sim p(\theta) whose simulations fall within a tolerance ϵ\epsilon of the observation, Dϵ(s(θ),xo)<ϵD_{\epsilon}(s(\theta),x_o) < \epsilon, usually after reducing the data to summary statistics [32]. It scales poorly with dim(x)(x) and is sensitive to the choice of summaries and of ϵ\epsilon. Neural SBI instead trains a density estimator qϕq_{\phi} on a corpus of simulations (θi,xi)(\theta_i,x_i) with θi∼p(θ)\theta_i \sim p(\theta) and xi∼p(x∣θi)x_i \sim p(x \mid\theta_i), targeting the posterior (NPE) [2, 3], the likelihood (NLE) [4] or the likelihood-to-evidence ratio (NRE) [5, 6, 20]. Amortised variants train once to learn p(θ)p(x∣θ)p(\theta)p(x \mid\theta) and apply to any xx, whereas sequential variants concentrate simulations around xox_o, giving up amortisation for simulation efficiency.

Diffusion Sampling

Neural diffusion samplers learn a Stochastic Differential Equation (SDE) which transports samples X0∼μ(x)X_0 \sim\mu(x) from source distribution μ(x)\mu(x); at time t=0t = 0 to the Boltzmann distribution v(x)v(x) at time t=1t = 1 given by:

v(x)=e−E(x)/TZ,Z=∫Xe−E(x)/T dx<∞(1)v(x) = \frac{e^{-E(x)/T}}{Z}, \qquad Z = \int_{\mathcal{X}} e^{-E(x)/T}\,dx < \infty \tag*{(1)}

The distribution is characterized by an unnormalized energy E:X⊆Rd→RE : \mathcal{X} \subseteq\mathbb{R}^{d} \to\mathbb{R} and temperature TT, with the normalizing constant ZZ being intractable. Diffusion samplers aim to provide faster convergence using fewer evaluations of the energy function than Markov Chain Monte Carlo (MCMC) [33] methods by learning better proposal distributions. Especially when v(x)v(x) is defined through an energy function, diffusion samplers provide a scalable alternative to sampling methods such as score matching [34, 35] or normalizing flows [2, 4] whose naive implementations rely on samples from v(x)v(x).

Adjoint Schrödinger Bridge Sampling: ASBS[14] is a state-of-the-art diffusion sampler building on Adjoint Sampling (AS)[15] that solves the Schrödinger Bridge (SB) problem between source distribution μ(x)\mu(x) and target v(x)v(x) through stochastic optimal control. Within the context of RSBI, we consider the following drift-less SDE:

dXt=σtut(Xt)dt+σtdWt,X0∼μ(X0),(2)dX_t = \sigma_t u_t(X_t)dt + \sigma_t dW_t,\qquad X_0 \sim\mu(X_0), \tag*{(2)}

where σt\sigma_t is the diffusion coefficient, utu_t is a learned control network, and (Wt)t(W_t)_t is a dd-dimensional Brownian motion. The optimal control ut∗u_t^* is then given by:

u∗=arg⁡min⁡u DKL(pu∥pbase)=arg⁡min⁡u EX∼pu[∫0112∥ut(Xt)∥2dt],(3)u^* = \underset{u}{\arg\min}\ D_{\mathrm{KL}}(p^u\|p^{\mathrm{base}}) = \underset{u}{\arg\min}\ \mathbb{E}_{X\sim p^u}\left[\int_0^1 \frac{1}{2}\|u_t(X_t)\|^2dt\right], \tag*{(3)}
s.t. dXt=σtut(Xt)dt+σtdWt,X0∼μ(X0),X1∼v(X1).(4)\text{s.t. } dX_t = \sigma_tu_t(X_t)dt + \sigma_tdW_t,\qquad X_0 \sim\mu(X_0),\qquad X_1 \sim v(X_1). \tag*{(4)}

Therefore, u∗u^* minimizes the relative entropy between the path probability pup^u characterized by (2) and the base path probability pbasep^{\mathrm{base}} given when u=0u = 0. To solve (3), ASBS defines an additional corrector network hh and iterates over two objectives; Adjoint Matching (AM):

u(k)=arg⁡min⁡u Ept∣0,1basep0,1u[∥ut(Xt)+σt(∇X1E(X1)+h(k−1)(X1))∥2],uˉ=stopgrad⁡(u),(5)u^{(k)} = \underset{u}{\arg\min}\ \mathbb{E}_{p^{\mathrm{base}}_{t|0,1}p^u_{0,1}}\left[\left\|u_t(X_t) + \sigma_t\left(\nabla_{X_1}E(X_1) + h^{(k-1)}(X_1)\right)\right\|^2\right],\qquad\bar{u} = \operatorname{stopgrad}(u), \tag*{(5)}

and Corrector Matching (CM):

References

  1. [1]Kyle Cranmer, Johann Brehmer, and Gilles Louppe. The frontier of simulation-based inference. Proceedings of the National Academy of Sciences, 117(48):30055–30062, 2020. doi: 10.1073/pnas.1912789117. URL https://www.pnas.org/doi/abs/10.1073/pnas.1912789117.
  2. [2]George Papamakarios and Iain Murray. Fast e-free inference of simulation models with bayesian conditional density estimation. In Daniel D. Lee, Masashi Sugiyama, Ulrike von Luxburg, Isabelle Guyon, and Roman Garnett, editors, Advances in Neural Information Processing Systems 29: Annual Conference on Neural Information Processing Systems 2016, December 5-10, 2016, Barcelona, Spain, pages 1028–1036, 2016. URL https://proceedings.neurips.cc/paper/2016/hash/6aca97005c68f1206823815f66102863-Abstract.html.
  3. [3]David S. Greenberg, Marcel Nonnenmacher, and Jakob H. Macke. Automatic posterior transformation for likelihood-free inference. In Kamalika Chaudhuri and Ruslan Salakhutdinov, editors, Proceedings of the 36th International Conference on Machine Learning, ICML 2019, 9-15 June 2019, Long Beach, California, USA, Proceedings of Machine Learning Research, pages 2404–2414. PMLR, 2019. URL http://proceedings.mlr.press/v97/greenberg19a.html.arxiv.org/abs/1905.07488
  4. [4]George Papamakarios, David C. Sterratt, and Iain Murray. Sequential neural likelihood: Fast likelihood-free inference with autoregressive flows. In Kamalika Chaudhuri and Masashi Sugiyama, editors, The 22nd International Conference on Artificial Intelligence and Statistics, AISTATS 2019, 16-18 April 2019, Naha, Okinawa, Japan, Proceedings of Machine Learning Research, pages 837–848. PMLR, 2019. URL http://proceedings.mlr.press/v89/papamakarios19a.html.arxiv.org/abs/1805.07226
  5. [5]Joeri Hermans, Volodimir Begy, and Gilles Louppe. Likelihood-free MCMC with amortized approximate ratio estimators. In Proceedings of the 37th International Conference on Machine Learning, ICML 2020, 13-18 July 2020, Virtual Event, Proceedings of Machine Learning Research, pages 4239–4248. PMLR, 2020. URL http://proceedings.mlr.press/v119/hermans20a.html.arxiv.org/abs/1903.04057
  6. [6]Conor Durkan, Iain Murray, and George Papamakarios. On contrastive learning for likelihood-free inference. In Proceedings of the 37th International Conference on Machine Learning, ICML 2020, 13-18 July 2020, Virtual Event, Proceedings of Machine Learning Research, pages 2771–2781. PMLR, 2020. URL http://proceedings.mlr.press/v119/durkan20a.html.arxiv.org/abs/2002.03712
  7. [7]Manuel Glöckler, Michael Deistler, and Jakob H. Macke. Variational methods for simulation-based inference. In The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022. OpenReview.net, 2022. URL https://openreview.net/forum?id=kZOUYdhqkNY.arxiv.org/abs/2203.04176
  8. [8]Benjamin Kurt Miller, Alex Cole, Patrick Forré, Gilles Louppe, and Christoph Weniger. Truncated marginal neural ratio estimation. In Advances in Neural Information Processing Systems, volume 34, 2021. URL https://papers.neurips.cc/paper_files/paper/2021/file/01632f7b7a127233fa188bd6c2e42e1-Paper.pdf.
  9. [9]Michael Deistler, Pedro J. Gonçalves, and Jakob H. Macke. Truncated proposals for scalable and hass-free simulation-based inference. In *Advances in Neural Information Processing Systems 35 (NeurIPS 2022)*, pages 23135–23149. Curran Associates, Inc., 2022. URL https://proceedings.neurips.cc/paper_files/paper/2022/hash/9278abf072b58caf2 1d48dd670b4c721-Abstract-Conference.html.DOI
  10. [10]Maximilian Dax, Stephen R. Green, Jonathan Gair, Nihar Gupte, Michael Pürrer, Vivien Raymond, Jonas Wildberger, Jakob H. Macke, Alessandra Buonanno, and Bernhard Schölkopf. Real-time inference for binary neutron star mergers using machine learning. Nature, 639:49–53, 2025. doi: 10.1038/s41586-025-08593-z. URL https://www.nature.com/articles/s4 1586-025-08593-z.DOI
  11. [11]Maximilian Dax, Stephen R. Green, Jonathan Gair, Michael Pürrer, Jonas Wildberger, Jakob H. Macke, Alessandra Buonanno, and Bernhard Schölkopf. Neural importance sampling for rapid and reliable gravitational-wave inference. Physical Review Letters, 130(17):171403, 2023. doi: 10.1103/PhysRevLett.130.171403. URL https://doi.org/10.1103/PhysRevLett.130. 171403.
  12. [12]James Bradbury, Roy Frostig, Peter Hawkins, Matthew James Johnson, Yash Katariya, Chris Leary, Dougal Maclaurin, George Necula, Adam Paszke, Jake VanderPlas, Skye Wanderman-Milne, and Qiao Zhang. JAX: composable transformations of Python+NumPy programs, 2018. URL http://github.com/jax-ml/jax.
  13. [13]Jason Ansel, Edward Yang, Horace He, Natalia Gimelshein, Animesh Jain, Michael Voznesensky, Bin Bao, Peter Bell, David Berard, Evgeni Burovski, Geeta Chauhan, Anjali Chourdia, Will Constable, Alban Desmaison, Zachary DeVito, Elias Ellison, Will Feng, Jiong Gong, Michael Gschwind, Brian Hirsh, Sherlock Huang, Kshiteej Kalambarkar, Laurent Kirsch, Michael Lazor, Mario Lezcano, Yanbo Liang, Jason Liang, Yinghai Lu, CK Luk, Bert Maher, Yunjie Pan, Christian Puhrsch, Matthias Reso, Mark Saroufim, Marcos Yukio Siraichi, Helen Suk, Michael Suo, Phil Tillet, Eikan Wang, Xiaodong Wang, William Wen, Shunting Zhang, Xu Zhao, Keren Zhou, Richard Zou, Ajit Mathews, Gregory Chanan, Peng Wu, and Soumith Chintala. PyTorch 2: Faster Machine Learning Through Dynamic Python Bytecode Transformation and Graph Compilation. In 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2 (ASPLOS ’24). ACM, April 2024. doi: 10.1145/3620665.3640366. URL https://docs.python.org/assets/python2-2.pdf.DOI
  14. [14]Guan-Horng Liu, Jaemoo Choi, Yongxin Chen, Benjamin Kurt Miller, and Ricky T. Q. Chen. Adjoint schrödinger bridge sampler. In Danielle Belgrave, Cheng Zhang, Laura N. Montoya, Hsuan-Tien Lin, Razvan Pascanu, Piotr Koniusz, Marzyeh Ghassemi, Nancy Chen, Iván Vladimir Meza Ruíz, and Arturo Loaiza-Bonilla, editors, Advances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing Systems 2025, NeurIPS 2025, San Diego, CA, USA, December 2-7, 2025 / Mexico City, Mexico, November 30 - December 5, 2025, 2025. URL http://papers.nips.cc/paper_files/paper/2025/ha sh/174692c52dc84fad2b2e99dd8637ce6a-Abstract-Conference.html.arxiv.org/abs/2506.22565
  15. [15]Aaron J. Havens, Benjamin Kurt Miller, Bing Yan, Carles Domingo-Enrich, Anuroop Sriram, Daniel S. Levine, Brandom M. Wood, Bin Hu, Brandon Amos, Brian Karner, Xiang Fu, Guan-Horng Liu, and Ricky T. Q. Chen. Adjoint sampling: Highly scalable diffusion samplers via adjoint matching. In Aarti Singh, Maryam Fazel, Daniel Hsu, Simon Lacoste-Julien, Felix Berkenkamp, Tegan Maharaj, Kiri Wagstaff, and Jerry Zhu, editors, Forty-second International Conference on Machine Learning, ICML 2025, Vancouver, BC, Canada, July 13-19, 2025, Proceedings of Machine Learning Research. PMLR, 2025. URL https://proceedings.ml r.press/v267/havens25a.html.arxiv.org/abs/2504.11713
  16. [16]Juno Nam, Bálint Máté, Artur P. Toshev, Manasa Kaniselvan, Rafael Gomez-Bombarelli, Ricky T. Q. Chen, Brandon M. Wood, Guan-Horng Liu, and Benjamin Kurt Miller. Enhancing diffusion-based sampling with molecular collective variables. In The Fourteenth International Conference on Learning Representations, 2026. URL https://openreview.net/forum?i d=1bJN1EQByS.arxiv.org/abs/2510.11923
  17. [17]Alessandro Barducci, Giovanni Bussi, and Michele Parrinello. Well-tempered metadynamics: A smoothly converging and tunable free-energy method. Phys. Rev. Lett., 100:020603, Jan 2008. doi: 10.1103/PhysRevLett.100.020603. URL https://link.aps.org/doi/10.1103/PhysRevLett.100.020603.DOI
  18. [18]Jan-Matthis Lueckmann, Jan Boelts, David Greenberg, Pedro Goncalves, and Jakob Macke. Benchmarking simulation-based inference. In Arindam Banerjee and Kenji Fukuzumi, editors, Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, volume 130 of Proceedings of Machine Learning Research, pages 343–351. PMLR, 13–15 Apr 2021. URL https://proceedings.mlr.press/v130/lueckmann21a.html.arxiv.org/abs/2101.04653
  19. [19]George Papamakarios, Eric T. Nalisnick, Danilo Jimenez Rezende, Shakir Mohamed, and Balaji Lakshminarayanan. Normalizing flows for probabilistic modeling and inference. J. Mach. Learn. Res., 22(57):1–64, 2021. URL https://jmlr.org/papers/v22/19-1028.html.
  20. [20]Benjamin Kurt Miller, Christoph Weniger, and Patrick Forré. Contrastive neural ratio estimation. In Sammi Koyejo, S. Mohamed, A. Agarwal, Danielle Belgrave, K. Cho, and A. Oh, editors, Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems 2022, NeurIPS 2022, New Orleans, LA, USA, November 28 - December 9, 2022, 2022. URL http://papers.nips.cc/paper_files/paper/2022/hash/159f7fe5b51ecd663b85337e8e28ec65-Abstract-Conference.html.DOI
  21. [21]Mark A Beaumont, Wenyang Zhang, and David J Balding. Approximate bayesian computation in population genetics. Genetics, 162(4):2025–2035, 12 2002. ISSN 1943-2631. doi: 10.1093/genetics/162.4.2025. URL https://doi.org/10.1093/genetics/162.4.2025.
  22. [22]Alexander Moreno, Tameem Adel, Edward Meeds, James M. Rehg, and Max Welling. Automatic variational abc, 2016. URL https://arxiv.org/abs/1606.08549.
  23. [23]Max Welling and Yee Whye Teh. Bayesian learning via stochastic gradient langevin dynamics. In Lise Getoor and Tobias Scheffer, editors, Proceedings of the 28th International Conference on Machine Learning, ICML 2011, Bellevue, Washington, USA, June 28 - July 2, 2011, pages 681–688. Omnipress, 2011. URL https://icml.cc/2011/papers/398_icmlpaper.pdf.
  24. [24]Lorenzo Pacchiardi, Sherman Khoo, and Ritabrata Dutta. Generalized Bayesian likelihood-free inference. Electronic Journal of Statistics, 18(2):3628 – 3686, 2024. doi: 10.1214/24-EJS2283. URL https://doi.org/10.1214/24-EJS2283.
  25. [25]Yin Li, Chirag Modi, Drew Jamieson, Yucheng Zhang, Libin Lu, Yu Feng, François Lanusse, and Leslie Greengard. Differentiable cosmological simulation with the adjoint method. The Astrophysical Journal Supplement Series, 270(2):36, feb 2024. doi: 10.3847/1538-4365/ad0ce7. URL https://doi.org/10.3847/1538-4365/ad0ce7.
  26. [26]Thomas D. P. Edwards, Kaze W. K. Wong, Kelvin K. H. Lam, Adam Coogan, Daniel Foreman-Mackey, Maximiliano Isi, and Aaron Zimmerman. Differentiable and hardware-accelerated waveforms for gravitational wave data analysis. Phys. Rev. D, 110:064028, Sep 2024. doi: 10.1103/PhysRevD.110.064028. URL https://link.aps.org/doi/10.1103/PhysRevD.110.064028.DOI
  27. [27]Henrik Christiansen, Takashi Maruyama, Federico Errica, Viktor Zaverkin, Makoto Takamoto, and Francesco Alesiani. Fast, modular, and differentiable framework for machine learning-enhanced molecular simulations. The Journal of Chemical Physics, 163(18):182501, 11 2025. ISSN 0021-9606. doi: 10.1063/5.0277356. URL https://doi.org/10.1063/5.0277356.
  28. [28]Satvik Mishra, Roberto Trotta, and Matteo Viel. Large, fast, and accurate H I intensity maps with latent overlap diffusion. Monthly Notices of the Royal Astronomical Society, 545(3):staf2071, January 2026. doi: 10.1093/mnras/staf2071. URL https://doi.org/10.1093/mnras/staf2071.
  29. [29]Johann Brehmer, Gilles Louppe, Juan Pavez, and Kyle Cranmer. Mining gold from implicit models to improve likelihood-free inference. Proceedings of the National Academy of Sciences, 117(10):5242–5249, 2020. doi: 10.1073/pnas.1915980117. URL https://www.pnas.org/doi/abs/10.1073/pnas.1915980117.DOI
  30. [30]Eric Jang, Shixiang Gu, and Ben Poole. Categorical reparameterization with gumbel-softmax. In 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings. OpenReview.net, 2017. URL https://openreview.net/forum?id=rkE3y85ee.arxiv.org/abs/1611.01144
  31. [31]Justine Zeghal, François Lanusse, Alexandre Boucaud, Benjamin Remy, and Eric Aubourg. Neural posterior estimation with differentiable simulators. In ICML 2022 Workshop on Machine Learning for Astrophysics, 2022.arxiv.org/abs/2207.05636
  32. [32]Jean-Michel Marin, Pierre Pudlo, Christian P. Robert, and Robin J. Ryder. Approximate bayesian computational methods. Statistics and Computing, 22(6):1167–1180, Nov 2012. ISSN 1573-1375. doi: 10.1007/s11222-011-9288-2. URL https://doi.org/10.1007/s11222-011-9288-2.
  33. [33]Nicholas Metropolis, Arianna W. Rosenbluth, Marshall N. Rosenbluth, Augusta H. Teller, and Edward Teller. Equation of state calculations by fast computing machines. The Journal of Chemical Physics, 21(6):1087–1092, 06 1953. ISSN 0021-9606. doi: 10.1063/1.1699114. URL https://doi.org/10.1063/1.1699114.
  34. [34]Valentin De Bortoli, Michael J. Hutchinson, Peter Wirnsberger, and Arnaud Doucet. Target score matching. CoRR, abs/2402.08667, 2024. doi: 10.48550/ARXIV.2402.08667. URL https://doi.org/10.48550/arXiv.2402.08667.
  35. [35]Angus Phillips, Hai-Dang Dau, Michael John Hutchinson, Valentin De Bortoli, George Deligiannidis, and Arnaud Doucet. Particle denoising diffusion sampler. In Ruslan Salakhutdinov, Zico Kolter, Katherine A. Heller, Adrian Weller, Nuria Oliver, Jonathan Scarlett, and Felix Berkenkamp, editors, Forty-first International Conference on Machine Learning, ICML 2024, Vienna, Austria, July 21-27, 2024, Proceedings of Machine Learning Research, pages 40688–40724. PMLR / OpenReview.net, 2024. URL https://proceedings.mlr.press/v235/phillips24a.html.arxiv.org/abs/2402.06320
  36. [36]S. Kullback. Probability densities with given marginals. The Annals of Mathematical Statistics, 39(4):1236–1243, 1968. ISSN 00034851, 21688990. URL http://www.jstor.org/stable/2239692.DOI
  37. [37]Maria L. Rizzo and Gábor J. Székely. Energy distance. WIREs Computational Statistics, 8(1):27–38, 2016. doi: https://doi.org/10.1002/wics.1375. URL https://wires.onlinelibrary.wiley.com/doi/abs/10.1002/wics.1375.
  38. [38]Arthur Gretton, Karsten M. Borgwardt, Malte J. Rasch, Bernhard Schölkopf, and Alexander J. Smola. A kernel two-sample test. J. Mach. Learn. Res., 13:723–773, 2012. URL https://dl.acm.org/doi/10.5555/2503308.2188410.DOI
  39. [39]Radford M. Neal. MCMC using Hamiltonian dynamics. In Steve Brooks, Andrew Gelman, Galin L. Jones, and Xiao-Li Meng, editors, Handbook of Markov Chain Monte Carlo, chapter 5, pages 113–162. Chapman & Hall/CRC, 2011.DOI
  40. [40]Matthew D. Hoffman and Andrew Gelman. The no-u-turn sampler: adaptively setting path lengths in hamiltonian monte carlo. J. Mach. Learn. Res., 15(1):1593–1623, 2014. URL https://dl.acm.org/doi/10.5555/2627435.2638586.
  41. [41]David Lopez-Paz and Maxime Oquab. Revisiting classifier two-sample tests. In 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings. OpenReview.net, 2017. URL https://openreview.net/forum?id=SJkXfE5xx.
  42. [42]Daniel Foreman-Mackey, David W. Hogg, Dustin Lang, and Jonathan Goodman. emcee: The mcmc hammer. Publications of the Astronomical Society of the Pacific, 125(925):306–312, feb 2013. doi: 10.1086/670067. URL https://doi.org/10.1086/670067.
  43. [43]P. G. Bissiri, C. C. Holmes, and S. G. Walker. A general framework for updating belief distributions. Journal of the Royal Statistical Society Series B: Statistical Methodology, 78(5):1103–1130, 11 2016. ISSN 1369-7412. doi: 10.1111/rssb.12158. URL https://doi.org/10.1111/rssb.12158.
  44. [44]Gareth O. Roberts and Jeffrey S. Rosenthal. Optimal scaling for various metropolis-hastings algorithms. Statistical Science, 16(4):351–367, 2001. ISSN 08834237. URL http://www.jstor.org/stable/3182776.DOI
  45. [45]Arnak S. Dalalyan. Theoretical guarantees for approximate sampling from smooth and log-concave densities. Journal of the Royal Statistical Society Series B: Statistical Methodology, 79 (3):651–676, 06 2017. ISSN 1369-7412. doi: 10.1111/rssb.12183. URL https://doi.org/10.1111/rssb.12183.
  46. [46]Heikki Haario, Eero Saksman, and Johanna Tamminen. An adaptive metropolis algorithm. Bernoulli, 7(2):223–242, 2001. ISSN 1350-7265.DOI
  47. [47]Bob Carpenter, Andrew Gelman, Matthew D. Hoffman, Daniel Lee, Ben Goodrich, Michael Betancourt, Marcus Brubaker, Jiqiang Guo, Peter Li, and Allen Riddell. Stan: A probabilistic programming language. Journal of Statistical Software, 76(1):1–32, 2017. doi: 10.18637/jss.v076.i01. URL https://www.jstatsoft.org/index.php/jss/article/view/v076i01.DOI
  48. [48]Mark Girolami and Ben Calderhead. Riemann manifold langevin and hamiltonian monte carlo methods. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 73(2):123–214, 2011. doi: 10.1111/j.1467-9868.2010.00765.x. URL https://rss.onlinelibrary.wiley.com/doi/abs/10.1111/j.1467-9868.2010.00765.x.DOI
  49. [49]Matthew Hoffman, Pavel Sountsov, Joshua V. Dillon, Ian Langmore, Dustin Tran, and Srinivas Vasudevan. Neutra-lizing bad geometry in hamiltonian monte carlo using neural transport, 2019. URL https://arxiv.org/abs/1903.03704.
  50. [50]Adam Perbost, Francis Bach, and Pierre Marion. Diffusion removes Langevin’s conditioning dependence: A sharp Gaussian analysis. arXiv preprint arXiv:2610.12052, 2026.arxiv.org/abs/2610.12052
  51. [51]Shadab Ahamed, Eshed Gal, Md Shahriar Rahim Siddiqui, Simon Ghyselincks, Moshe Eliasof, and Eldad Haber. Preconditioned flow matching. In Advances in Neural Information Processing Systems, volume 39, 2026. URL https://arxiv.org/abs/2603.02337. To appear. Preprint version: arXiv:2603.02337.
  52. [52]Jonathan Goodman and Jonathan Weare. Ensemble samplers with affine invariance. Communications in Applied Mathematics and Computational Science, 5(1):65–80, 2010. doi: 10.2140/camcos.2010.5.65.DOI
  53. [53]Alvaro Tejero-Cantero, Jan Boelts, Michael Deistler, Jan-Matthis Lueckmann, Conor Durkan, Pedro J. Goncalves, David S. Greenberg, and Jakob H. Macke. sbi: A toolkit for simulation-based inference. Journal of Open Source Software, 5(52):2505, 2020. doi: 10.21105/joss.02505. URL https://doi.org/10.21105/joss.02505.
  54. [54]George Papamakarios, Iain Murray, and Theo Pavlakou. Masked autoregressive flow for density estimation. In Isabelle Guyon, Ulrike von Luxburg, Samy Bengio, Hanna M. Wallach, Rob Fergus, S. V. N. Vishwanathan, and Roman Garnett, editors, Advances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2017, December 4-9, 2017, Long Beach, CA, USA, pages 2338–2347, 2017. URL https://proceedings.neurips.cc/paper/2017/hash/6c1da886822c67822bcf3679d04369fa-Abstract.html.arxiv.org/abs/1705.07057
  55. [55]Conor Durkan, Artur Bekasov, Iain Murray, and George Papamakarios. Neural spline flows. In Hanna M. Wallach, Hugo Larochelle, Alina Beygelzimer, Florence d’Alché-Buc, Emily B. Fox, and Roman Garnett, editors, Advances in Neural Information Processing Systems 32: Annual Conference on Neural Information Processing Systems 2019, NeurIPS 2019, December 8-14, 2019, Vancouver, BC, Canada, pages 7509–7520, 2019. URL https://proceedings.neurips.cc/paper/2019/hash/7ac71d433f282034e088473244df8c02-Abstract.html.arxiv.org/abs/1906.04032
  56. [56]Paul Fearnhead and Dennis Prangle. Constructing summary statistics for approximate Bayesian computation: semi-automatic approximate Bayesian computation. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 74(3):419–474, 2012. doi: 10.1111/j.1467-9868.2011.01010.x. URL https://doi.org/10.1111/j.1467-9868.2011.01010.x.
  57. [57]Simon N. Wood. Statistical inference for noisy nonlinear ecological dynamic systems. Nature, 466(7310):1102–1104, 2010. doi: 10.1038/nature09319. URL https://doi.org/10.1038/nature09319.
  58. [58]Leah F. Price, Christopher C. Drovandi, Anthony Lee, and David J. Nott. Bayesian synthetic likelihood. Journal of Computational and Graphical Statistics, 27(1):1–11, 2018. doi: 10.1080/10618600.2017.1302882. URL https://doi.org/10.1080/10618600.2017.1302882.
  59. [59]Espen Bernton, Pierre E. Jacob, Mathieu Gerber, and Christian P. Robert. Approximate Bayesian computation with the Wasserstein distance. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 81(2):235–269, 2019. doi: 10.1111/rssb.12312. URL https://doi.org/10.1111/rssb.12312.
  60. [60]Mijung Park, Wittawat Jitkrittum, and Dino Sejdinovic. K2-abc: Approximate bayesian computation with kernel embeddings. In Arthur Gretton and Christian C. Robert, editors, Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, volume 51 of Proceedings of Machine Learning Research, pages 398–407, Cadiz, Spain, 09–11 May 2016. PMLR. URL https://proceedings.mlr.press/v51/park16.html.arxiv.org/abs/1502.02558
  61. [61]Bai Jiang. Approximate Bayesian computation with Kullback-Leibler divergence as data discrepancy. In Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics, volume 84 of Proceedings of Machine Learning Research, pages 1711–1721, 2018. URL https://proceedings.mlr.press/v84/jiang18a.html.
  62. [62]Hien Duy Nguyen, Julyan Arbel, Hongliang Lü, and Florence Forbes. Approximate Bayesian computation via the energy statistic. IEEE Access, 8:131683–131698, 2020. doi: 10.1109/ACCESS.2020.3009878. URL https://doi.org/10.1109/ACCESS.2020.3009878.
  63. [63]Christopher Drovandi and David T. Frazier. A comparison of likelihood-free methods with and without summary statistics. Statistics and Computing, 32(3):42, 2022. doi: 10.1007/s11222-022-10092-4. URL https://doi.org/10.1007/s11222-022-10092-4.
  64. [64]Richard David Wilkinson. Approximate Bayesian computation (ABC) gives exact results under the assumption of model error. Statistical Applications in Genetics and Molecular Biology, 12(2):129–141, 2013. doi: 10.1515/sagmb-2013-0010. URL https://doi.org/10.1515/sagmb-2013-0010.
  65. [65]David A. Nix and Andreas S. Weigend. Estimating the mean and variance of the target probability distribution. In Proceedings of 1994 IEEE International Conference on Neural Networks (ICNN’94), volume 1, pages 55–60. IEEE, 1994. doi: 10.1109/ICNN.1994.374138. URL https://ieeexplore.ieee.org/document/374138/.
  66. [66]Balaji Lakshminarayanan, Alexander Pritzel, and Charles Blundell. Simple and scalable predictive uncertainty estimation using deep ensembles. In Advances in Neural Information Processing Systems 30 (NIPS 2017), 2017. URL https://proceedings.neurips.cc/paper/2017/hash/9ef2edb4f2dc810847ffa5fa856bce38-Abstract.html.arxiv.org/abs/1612.01474
  67. [67]Gábor J. Székely and Maria L. Rizzo. Energy statistics: A class of statistics based on distances. Journal of Statistical Planning and Inference, 143(8):1249–1272, 2013. ISSN 0378-3758. doi: 10.1016/j.jspi.2013.03.018. URL https://www.sciencedirect.com/science/article/pii/S0378375813000633.DOI
  68. [68]Dino Sejdinovic, Bharath Sriperumbudur, Arthur Gretton, and Kenji Fukumizu. Equivalence of distance-based and RKHS-based statistics in hypothesis testing. The Annals of Statistics, 41(5):2263–2291, 2013. doi: 10.1214/13-AOS1140. URL https://projecteuclid.org/journals/annals-of-statistics/volume-41/issue-5/Equivalence-of-distance-based-and-RKHS-based-statistics-in-hypothesis/10.1214/13-AOS1140.full.DOI
  69. [69]Michael G. B. Blum. Approximate bayesian computation: A nonparametric perspective. Journal of the American Statistical Association, 105(491):1178–1187, 2010. ISSN 01621459. URL http://www.jstor.org/stable/27920142.arxiv.org/abs/0904.0635
  70. [70]Marco Crisostomi, Kallol Dey, Enrico Barausse, and Roberto Trotta. Neural posterior estimation with guaranteed exact coverage: The ringdown of gw150914. Phys. Rev. D, 108:044029, Aug 2023. doi: 10.1103/PhysRevD.108.044029. URL https://link.aps.org/doi/10.1103/PhysRevD.108.044029.DOI
  71. [71]Abhirup Ghosh, Nathan K. Johnson-McDaniel, Archisman Ghosh, Chandra Kant Mishra, Parameswaran Ajith, Walter Del Pozzo, Christopher P. L. Berry, Alex B. Nielsen, and Lionel London. Testing general relativity using gravitational wave signals from the inspiral, merger and ringdown of binary black holes. Classical and Quantum Gravity, 35(1):014002, January 2018. doi: 10.1088/1361-6382/aa972e.DOI
  72. [72]Swetha Bhagwat, Xisco Jiménez Forteza, Paolo Pani, and Valeria Ferrari. Ringdown overtones, black hole spectroscopy, and no-hair theorem tests. Phys. Rev. D, 101:044033, Feb 2020. doi: 10.1103/PhysRevD.101.044033. URL https://link.aps.org/doi/10.1103/PhysRevD.101.044033.DOI
  73. [73]Tero Karras, Miika Aittala, Timo Aila, and Samuli Laine. Elucidating the design space of diffusion-based generative models. In Sanmi Koyejо, S. Mohamed, A. Agarwal, Danielle Belgrave, K. Cho, and A. Oh, editors, Advances in Neural Information Processing Systems 35: Annual Conference on Neural Information Processing Systems 2022, NeurIPS 2022, New Orleans, LA, USA, November 28 - December 9, 2022, 2022. URL http://papers.nips.cc/paper_files/paper/2022/hash/a98846e9d9c01cfb87eb694d946ce6b-Abstract-Conference.html.arxiv.org/abs/2206.00364
  74. [74]Luke Metz, C. Daniel Freeman, Samuel S. Schoenholz, and Tal Kachman. Gradients are not all you need. CoRR, abs/2111.05803, 2021. doi: 10.48550/ARXIV.2111.05803. URL https://doi.org/10.48550/arXiv.2111.05803.
  75. [75]Yoshua Bengio, Jérôme Louradour, Ronan Collobert, and Jason Weston. Curriculum learning. In Andrea Pohoreckyj Danyluk, Léon Bottou, and Michael L. Littman, editors, Proceedings of the 26th Annual International Conference on Machine Learning, ICML 2009, Montreal, Quebec, Canada, June 14-18, 2009, volume 382 of ACM International Conference Proceeding Series, pages 41–48. ACM, 2009. doi: 10.1145/1553374.1553380. URL https://doi.org/10.1145/1553374.1553380.
  76. [76]H.G. Bock and K.J. Plitt. A multiple shooting algorithm for direct solution of optimal control problems. IFAC Proceedings Volumes, 17(2):1603–1608, 1984. ISSN 1474-6670. doi: 10.1016/S1474-6670(17)61205-9. URL https://www.sciencedirect.com/science/article/pii/S1474667017612059. 9th IFAC World Congress: A Bridge Between Control Science and Technology, Budapest, Hungary, 2-6 July 1984.DOI
  77. [77]Henning U. Voss, Jens Timmer, and Jürgen Kurths. Nonlinear dynamical system identification from uncertain and indirect measurements. Int. J. Bifurc. Chaos, 14(6):1905–1933, 2004. doi: 10.1142/S0218127404010345. URL https://doi.org/10.1142/S0218127404010345.
  78. [78]Evren Mert Turan and Johannes Jäschke. Multiple shooting for training neural differential equations on time series. IEEE Control. Syst. Lett., 6:1897–1902, 2022. doi: 10.1109/LCSYS.2021.3135835. URL https://doi.org/10.1109/LCSYS.2021.3135835.
  79. [79]J. O. Ramsay, G. Hooker, D. Campbell, and J. Cao. Parameter estimation for differential equations: a generalized smoothing approach. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 69(5):741–796, 2007. doi: 10.1111/j.1467-9868.2007.00610.x. URL https://rss.onlinelibrary.wiley.com/doi/abs/10.1111/j.1467-9868.2007.00610.x.DOI
  80. [80]Nicholas G. Polson, James G. Scott, and Jesse Windle. Bayesian inference for logistic models using pólya–gamma latent variables. Journal of the American Statistical Association, 108(504):1339–1349, 2013. doi: 10.1080/01621459.2013.829001. URL https://doi.org/10.1080/01621459.2013.829001.

Paper details

vlnc:W_4s5g3xfj
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 10

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

Not recorded for this paper

View statement
Looking for related papers…
Contents