4–7 July 2026 • Tokyo, Japan

Information Geometry, Privacy and Monte Carlo

ISBA Satellite Meeting 2026

View Programme ISBA 2026

Key Dates

  • Poster proposals: 31 Jan 2026, 23:59 JST
  • Notification of acceptance: 20 Feb 2026
  • Registration deadline: Closed
  • Conference: July 4–7

About

An ISBA Satellite of the 2026 ISBA World Meeting, hosted at and supported by the Institute of Statistical Mathematics, Tokyo. The meeting brings together researchers and practitioners in information geometry, privacy, and modern Monte Carlo—including PDMP samplers, advanced MCMC, and variational methods.

  • 4-day programme: Sat 4 – Tue 7 July 2026 (Sat afternoon start)
  • 25 invited talks (35 min) + two evening poster sessions (~30 posters total)
  • Venue: ISM Main Conference Hall (capacity ~100); expected 60–80 participants
  • Registration: Free (pending funding confirmation)

Quick facts

  • Institute of Statistical Mathematics, 10-3 Midori-cho, Tachikawa, Tokyo 190-8562, Japan.
  • Nearest station: Takamatsu (Tama Monorail), ~10 min walk
  • From Tokyo Station: ~55–65 min by train (JR + Tama Monorail)

Registration

Participation is free of charge; however, advance online registration is required. Registration is closed.

  • Registration deadline: Closed
  • Capacity: approx. 80 participants (first-come, first-served)

Notes

  • We may close registration early if capacity is reached.

Programme

Abstract Booklet (PDF)

Provisional programme. Abstracts will be provided separately in the PDF booklet.

Day 1 — Sat, July 4

Afternoon only

12:30–13:00 Registration
13:00–13:15 Opening remarks
Chair: A. Beskos
13:15–13:50 Ajay JasraParticle Filtering for a Class of State-Space Models with Low and Degenerate Observational NoiseCUHK-Shenzhen
13:50–14:25 Christophe AndrieuThe Oracle Metropolis-HastingsUniversity of Bristol, UK
14:25–14:40 Coffee break
Chair: A. Beskos
14:40–15:15 Maria De IorioTitle TBATBA
15:15–15:50 Takuya KoriyamaAlignment from Pairwise Preferences: A Markov Chain ApproachUniversity of Chicago
15:50–16:25 Alexandros BeskosParticle Based Inference for Continuous-Discrete State Space ModelsUniversity College London
16:45–18:30 Poster Session (1st)

Day 2 — Sun, July 5

Full day

Chair: S. Mano
09:00–09:35 Federica MilinanniRapid mixing of stereographic MCMC for heavy-tailed samplingNorthwestern University
09:35–10:10 Kenji FukumizuTitle TBAThe Institute of Statistical Mathematics
10:10–10:25 Coffee break
Chair: C. P. Robert
10:25–11:00 Hongsheng DaiOnline federated learning framework for classificationNewcastle University
11:00–11:35 Stefano FavaroBayesian nonparametric privacy-preserving synthetic data generationUniversity of Torino and Collegio Carlo Alberto
11:35–12:10 Joshua BonPersuasive PrivacyAdelaide University
12:10–13:30 Lunch
13:30–13:50 Group photo
13:50–15:20 Poster Session (2nd)

Day 3 — Mon, July 6

Full day

Chair: D. Vats
09:00–09:35 Radu V. CraiuMCMC for Directed Acyclic Graphs via Birth-Death ProcessesUniversity of Toronto
09:35–10:10 Galin JonesTitle TBAUniversity of Minnesota
10:10–10:25 Coffee break
Chair: D. Vats
10:25–11:00 James FlegalSimultaneous confidence bands for (Markov chain) Monte Carlo simulationsUniversity of California, Riverside
11:00–11:35 Michiko OkudoApplications of information geometry to Bayesian prediction and estimation in curved exponential familiesChiba University
11:35–12:10 Marta CatalanoDistances on random measures for Bayesian nonparametricsLuiss University
12:10–13:30 Lunch
Chair: K. Yano
13:30–14:05 Takemasa MiyoshiTitle TBARIKEN
14:05–14:40 Manon MichelLatent Monte CarloCNRS, Universite Clermont-Auvergne
14:40–14:55 Coffee break
Chair: M. Choi
14:55–15:30 Geoffrey WolferCharacterization of Exponential Families of Lumpable Stochastic MatricesTokyo University of Agriculture and Technology
15:30–16:05 Daniel PaulinStochastic gradient Langevin dynamics: convergence and biasNanyang Technological University
16:05–16:40 Shahab AsoodehRecent Advances in Metropolis-Hastings AlgorithmsMcMaster University and Vector Institute

Day 4 — Tue, July 7

Morning only

Chair: M. Choi
09:00–09:35 Weiming FengFaster mixing of the Jerrum-Sinclair chainThe University of Hong Kong
09:35–10:10 Michael ChoiOptimising two-block averaging kernels to speed up Markov chainsNational University of Singapore
10:10–10:25 Coffee break
Chair: K. Yano
10:25–11:00 Yuga IguchiDynamical Regimes of Denoising Diffusion Models: Geometric PerspectivesSchool of Mathematical Sciences, Lancaster University
11:00–11:35 Edric TamSome Fundamental Limits to Neural Monte CarloStanford University
11:35–12:10 Victor ElviraTitle TBATBA
12:10–12:30 Closing and general discussion
12:30 Workshop ends

Posters

Sanket AgrawalUniversity of WarwickExit times for the Zig-Zag process in the presence of outlier minimodes
Daniel AndradeHiroshima UniversityReference posterior predictive distributions for benchmarking inference in Bayesian neural networks
Apratim ShuklaIndian Institute of Technology KanpurProximal Hamiltonian Monte Carlo
Yamato AraiUniversity of TokyoBridging Mixing and Fidelity in Continuous-Relaxation MCMC for Discrete Distributions
Youngsoo BaekDuke UniversitySafe, Scalable and Calibrated Stochastic Gradient MCMC for Hierarchical GLMs
Laura BazahicaLUT UniversityEfficient Amortized Bayesian Inference for Markov Random Fields via Gradient-Informed Grid Selection
Stefano CabrasUC3M, SpainBayesian frameworks for explicit biological age estimation
El Houssaine ChahbounUniversity of WarwickUnderstanding VR-IWAE: Gradient Reliability and Error Bounds
David ChenNational University of SingaporeScalable density regression using logistic Gaussian processes: a generalized variational approach
Yuan ChenUniversity of BristolThe Directed Gibbs Sampler
Maria Yus Trinity IrsanPresident UniversityCredibility frequency of zero-inflated Poisson distribution via variational Bayes approximation
Gonzalo García-DonatoUCLMSearching in ultra high dimensional sparse model spaces: New performance tests and boosting Gibbs sampling algorithms
Shufei GeShanghaiTech UniversityInformation-borrowed Bayesian model for (ultra) high-dimensional brain connectivity inference
Sebastiano GrazziBocconi UniversitySub-Cauchy Sampler
Luke HardcastleMRC Biostatistics Unit, University of CambridgeSticky manifold PDMPs
Max HirdUniversity of WaterlooNon-asymptotic Guarantees for Preconditioned MCMC under a Wasserstein-2 Contraction
Hsin-Hsiung HuangUniversity of Central FloridaBayesian Multilayer Mixed-Membership Networks: Fisher Information Guided Gaussian Variational Inference for Uncertainty Quantification
Jack JewsonMonash UniversityDifferentially Private Statistical Inference through beta-Divergence One Posterior Sampling
Mohamed Mehdi KetebESSECUnbiased Estimation of Log Normalizing Constants via Coupling and Path Sampling
El Mahdi KhribchESSEC Business SchoolOn the bias of Importance Sampling and Independent Metropolis-Hastings
Youngwoo KwonUniversity of MinnesotaSolidarity of Spectral Gaps for Component-Wise Markov Chains
Larissa LemosFGVTractable Riemann-Laplace Approximations
Antoine LucianoUniversité Paris-DauphineBayesian Adversarial Privacy
Minh Long NguyenQueensland University of TechnologyMoE-ZVCV: Mixture-of-Experts Zero-Variance Control Variates for Monte Carlo Integration
Yoshiyuki NinomiyaThe Institute of Statistical MathematicsPrior-intensified information criterion for spatially varying coefficient models
Lorenzo RimellaUniversity of BergamoScalable calibration of individual-based epidemic models through categorical approximations
Lucas SchwengberUC BerkeleyPosterior local sensitivity using Otto's calculus
Ruben SeyerChalmers/GURebalancing to Non-Reversible Continuous-Time Jump Samplers
Hirofumi ShibaThe Institute of Statistical MathematicsA PDMP Scaling Analysis: Gains from Persistent Momentum
Siva SivaganesanUniversity of CincinnatiBayesian Multiple Testing of One-Sided Hypotheses of two Proportions using Default Priors and an Approximate formulation of Hypotheses
Shijia WangShanghaiTech UniversityA multifidelity approximate Bayesian computation with adaptive pre-filtering sequential Monte Carlo
Zhihao WangUniversity of CopenhagenCouplings of stereographic MCMC algorithms
Yuyan WangNational University of SingaporeRobust Bayesian methods using amortized simulation-based inference

Venue

  • Venue: Institute of Statistical Mathematics, Tachikawa, Tokyo, Japan
  • Expected participants: 60–80
  • Format: In-person (no hybrid planned)

Hosts & Funding

Hosted by:The Institute of Statistical Mathematics

Supported by: The Institute of Statistical Mathematics and JST CREST (JPMJCR2115)

Organised by: Kengo Kamatani, Shuhei Mano, Keisuke Yano, Christian P. Robert, Alex Beskos, Michael Choi, Dootika Vats