mlsanityData to 7 Oct 2026

SIXO: Smoothing Inference with Twisted Objectives

arXiv 2206.05952 ↗First posted 13 Jun 2022cs.LG cs.AI stat.MLcited by 22

Not on the map. Why?

Abstract (arXiv)

Sequential Monte Carlo (SMC) is an inference algorithm for state space models that approximates the posterior by sampling from a sequence of target distributions. The target distributions are often chosen to be the filtering distributions, but these ignore information from future observations, leading to practical and theoretical limitations in inference and model learning. We introduce SIXO, a method that instead learns targets that approximate the smoothing distributions, incorporating information from all observations. The key idea is to use density ratio estimation to fit functions that warp the filtering distributions into the smoothing distributions. We then use SMC with these learned targets to define a variational objective for model and proposal learning. SIXO yields provably tighter log marginal lower bounds and offers significantly more accurate posterior inferences and parameter estimates in a variety of domains.

Key facts. Cited by 22 papers in the map (data through 7 Oct 2026). Builds on 17 of its 55 references that are in the map. First posted on arXiv 13 Jun 2022.

Builds on

17 of 55 references.

References · 17

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2013Auto-Encoding Variational Bayescited by 3,980
  3. 2016Variational Inference: A Review for Statisticianscited by 1,466
  4. 2015Importance Weighted Autoencoderscited by 454
  5. 2013Black Box Variational Inferencecited by 317
  6. 2016Learning in Implicit Generative Modelscited by 124
  7. 2017Variational Sequential Monte Carlocited by 93
  8. 2017Filtering Variational Objectivescited by 91
  9. 2016Variational inference for Monte Carlo objectivescited by 83
  10. 2017Auto-Encoding Sequential Monte Carlocited by 64
  11. 2019Elements of Sequential Monte Carlocited by 60
  12. 2015Neural Adaptive Sequential Monte Carlocited by 34
5 more
  1. 2019Energy-Inspired Models: Learning with Sampler-Induced Distributionscited by 12
  2. 2013Lookahead Strategies for Sequential Monte Carlocited by 5
  3. 2017Inference on high-dimensional implicit dynamic models using a guided intermediate resampling filtercited by 3
  4. 2016Particle Smoothing for Hidden Diffusion Processes: Adaptive Path Integral Smoothercited by 2
  5. 2019Graphical model inference: Sequential Monte Carlo meets deterministic approximationscited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 3; 2024: 7; 2025: 4; 2026: 8.

22 papers.

Sort:

By area:

  1. 2023Score-based Data Assimilationcited by 91
  2. 2024Particle Denoising Diffusion Samplercited by 63
  3. 2025Syntactic and Semantic Control of Large Language Models via Sequential Monte Carlocited by 45
  4. 2024Towards Scalable and Stable Parallelization of Nonlinear RNNscited by 38
  5. 2025Fast Controlled Generation from Language Models with Adaptive Weighted Rejection Samplingcited by 20
  6. 2024Inferring stochastic low-rank recurrent neural networks from neural datacited by 15
  7. 2024Bayesian Optimization of Antibodies Informed by a Generative Model of Evolving Sequencescited by 10
  8. 2024Amortized Control of Continuous State Space Feynman-Kac Model for Irregular Time Seriescited by 9
  9. 2024eXponential FAmily Dynamical Systems (XFADS): Large-scale nonlinear Gaussian state-space modelingcited by 7
  10. 2025Reinforced sequential Monte Carlo for amortised samplingcited by 5
  11. 2023NAS-X: Neural Adaptive Smoothing via Twistingcited by 4
  12. 2026On the Power of (Approximate) Reward Models for Inference-Time Scalingcited by 4
10 more
  1. 2026Contrastive Distribution Matching for Amortized Sequential Monte Carlo in Discrete Diffusioncited by 3
  2. 2026When are likely answers right? On Sequence Probability and Correctness in LLMscited by 3
  3. 2023Resampling Gradients Vanish in Differentiable Sequential Monte Carlo Samplerscited by 2
  4. 2026Reject, Resample, Repeat: Understanding Parallel Reasoning in Language Model Inferencecited by 2
  5. 2024Robust Inverse Graphics via Probabilistic Inferencecited by 1
  6. 2026FluxLite: Inference-Time Proposal Control for Discrete Diffusion Modelscited by 1
  7. 2025Efficient Inference for Coupled Hidden Markov Models in Continuous Time and Discrete Spacecited by 0
  8. 2026Neural Backward Filtering Forward Guidingcited by 0
  9. 2026Deep ZakaiJ: Structured Filtering for Jump-Diffusion Time Series Forecastingcited by 0
  10. 2026PR-Smoother: Simulator-Preserving Non-Gaussian Smoothing for Data Assimilationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2022Amortized backward variational inference in nonlinear state-space models
  2. 2024Probabilistic Inference in Language Models via Twisted Sequential Monte Carlo
  3. 2021Variational Rejection Particle Filtering
  4. 2026Inference-Time Alignment of Diffusion Models via Trust-Region Iterative Twisted Sequential Monte Carlo
  5. 2026Efficient Online Variational Estimation via Monte Carlo Sampling
  6. 2021Self-Supervised Inference in State-Space Models
  7. 2026Efficient Learning of Deep State Space Models via Importance Smoothing
  8. 2026Non-Asymptotic Error Bounds for SMC with Biased Proposals: Application to Conditional Diffusion Sampling