Improving Gradient-Guided Nested Sampling for Posterior Inference
Pablo Lemos, Nikolay Malkin, Will Handley, Yoshua Bengio, Yashar Hezaveh, Laurence Perreault Levasseur
Abstract
We present a performant, general-purpose gradient-guided nested sampling algorithm, , combining the state of the art in differentiable programming, Hamiltonian slice sampling, clustering, mode separation, dynamic nested sampling, and parallelization. This unique combination allows to scale well with dimensionality and perform competitively on a variety of synthetic and real-world problems. We also show the potential of combining nested sampling with generative flow networks to obtain large amounts of high-quality samples from the posterior distribution. This combination leads to faster mode discovery and more accurate estimates of the partition function.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
5 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- Iterated Denoising Energy Matching for Sampling from Boltzmann Densities cites Denoising Diffusion Probabilistic Models
- Adaptive teachers for amortized samplers cites Path Integral Sampler: A Stochastic Control Approach For Sampling
- Adaptive teachers for amortized samplers cites Generative Flow Networks for Discrete Probabilistic Modeling
- Adaptive teachers for amortized samplers cites Denoising Diffusion Probabilistic Models
- Adaptive teachers for amortized samplers cites Improved off-policy training of diffusion samplers
- Adaptive teachers for amortized samplers cites Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation
- Generative Flow Networks for Discrete Probabilistic Modeling cites Trajectory balance: Improved credit assignment in GFlowNets
- Trajectory balance: Improved credit assignment in GFlowNets cites Generative Flow Networks for Discrete Probabilistic Modeling
- Generative Flow Networks for Discrete Probabilistic Modeling cites Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation
- PQMass: Probabilistic Assessment of the Quality of Generative Models using Probability Mass Estimation cites Denoising Diffusion Probabilistic Models
- On scalable and efficient training of diffusion samplers cites Path Integral Sampler: A Stochastic Control Approach For Sampling
- On scalable and efficient training of diffusion samplers cites Iterated Denoising Energy Matching for Sampling from Boltzmann Densities
- On scalable and efficient training of diffusion samplers cites Adaptive teachers for amortized samplers
- On scalable and efficient training of diffusion samplers cites Improved off-policy training of diffusion samplers
- On scalable and efficient training of diffusion samplers cites Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation
- Trajectory balance: Improved credit assignment in GFlowNets cites Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation
- Improved off-policy training of diffusion samplers cites Path Integral Sampler: A Stochastic Control Approach For Sampling
- Improved off-policy training of diffusion samplers cites Generative Flow Networks for Discrete Probabilistic Modeling
- Improved off-policy training of diffusion samplers cites Denoising Diffusion Probabilistic Models
- Improved off-policy training of diffusion samplers cites Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation
Cited by top-tier papers5
- Iterated Denoising Energy Matching for Sampling from Boltzmann DensitiesTara Akhound-Sadegh, Jarrid Rector-Brooks, Avishek Joey Bose, Sarthak Mittal et al.ICML 2024 · 109 citations
- Improved off-policy training of diffusion samplersMarcin Sendera, Minsu Kim, Sarthak Mittal, Pablo Lemos et al.NeurIPS 2024 · 52 citations
- On scalable and efficient training of diffusion samplersMinkyu Kim, Kiyoung Seong, Dongyeop Woo, Sungsoo Ahn et al.NeurIPS 2025 · 11 citations
- Adaptive teachers for amortized samplersMinsu Kim, Sanghyeok Choi, Taeyoung Yun, Emmanuel Bengio et al.ICLR 2025
- PQMass: Probabilistic Assessment of the Quality of Generative Models using Probability Mass EstimationPablo Lemos, Sammy Nasser Sharief, Nikolay Malkin, Salma Salhi et al.ICLR 2025
Builds on7
- Denoising Diffusion Probabilistic ModelsJonathan Ho, Ajay Jain, Pieter AbbeelNeurIPS 2020 · 35,902 citations
- Flow Network based Generative Models for Non-Iterative Diverse Candidate GenerationEmmanuel Bengio, Moksh Jain, Maksym Korablyov, Doina Precup et al.NeurIPS 2021 · 565 citations
- Trajectory balance: Improved credit assignment in GFlowNetsNikolay Malkin, Moksh Jain, Emmanuel Bengio, Chen Sun et al.NeurIPS 2022 · 316 citations
- Path Integral Sampler: A Stochastic Control Approach For SamplingQinsheng Zhang, Yongxin ChenICLR 2022 · 177 citations
- Generative Flow Networks for Discrete Probabilistic ModelingDinghuai Zhang, Nikolay Malkin, Zhen Liu, Alexandra Volokhova et al.ICML 2022 · 131 citations
Related papers
- Deterministic Langevin Monte Carlo with Normalizing Flows for Bayesian InferenceRichard D. P. Grumitt, Biwei Dai, Uros SeljakNeurIPS 2022 · 15 citations
- QGFN: Controllable Greediness with Action ValuesElaine Lau, Stephen Zhewen Lu, Ling Pan, Doina Precup et al.NeurIPS 2024 · 21 citations
- Diffusive Gibbs SamplingWenlin Chen, Mingtian Zhang, Brooks Paige, José Miguel Hernández-Lobato et al.ICML 2024 · 21 citations
- Local Search GFlowNetsMinsu Kim, Taeyoung Yun, Emmanuel Bengio, Dinghuai Zhang et al.ICLR 2024 · 59 citations
- PhyloGFN: Phylogenetic inference with generative flow networksMing-Yang Zhou, Zichao Yan, Elliot Layne, Nikolay Malkin et al.ICLR 2024 · 32 citations
