This document summarizes dependent processes in Bayesian nonparametrics. It motivates the need for dependent random probability measures to accommodate temporal dependence structures beyond the exchangeability assumption. It describes modeling collections of random probability measures indexed by time as either discrete-time or continuous-time processes. The diffusive Dirichlet process is introduced as a dependent Dirichlet process with Dirichlet marginal distributions at each time point and continuous sample paths. Simulation and estimation methods are discussed for this model.
In this article we consider macrocanonical models for texture synthesis. In these models samples are generated given an input texture image and a set of features which should be matched in expectation. It is known that if the images are quantized, macrocanonical models are given by Gibbs measures, using the maximum entropy principle. We study conditions under which this result extends to real-valued images. If these conditions hold, finding a macrocanonical model amounts to minimizing a convex function and sampling from an associated Gibbs measure. We analyze an algorithm which alternates between sampling and minimizing. We present experiments with neural network features and study the drawbacks and advantages of using this sampling scheme.
In this article we consider macrocanonical models for texture synthesis. In these models samples are generated given an input texture image and a set of features which should be matched in expectation. It is known that if the images are quantized, macrocanonical models are given by Gibbs measures, using the maximum entropy principle. We study conditions under which this result extends to real-valued images. If these conditions hold, finding a macrocanonical model amounts to minimizing a convex function and sampling from an associated Gibbs measure. We analyze an algorithm which alternates between sampling and minimizing. We present experiments with neural network features and study the drawbacks and advantages of using this sampling scheme.
Bayesian inference for mixed-effects models driven by SDEs and other stochast...Umberto Picchini
An important, and well studied, class of stochastic models is given by stochastic differential equations (SDEs). In this talk, we consider Bayesian inference based on measurements from several individuals, to provide inference at the "population level" using mixed-effects modelling. We consider the case where dynamics are expressed via SDEs or other stochastic (Markovian) models. Stochastic differential equation mixed-effects models (SDEMEMs) are flexible hierarchical models that account for (i) the intrinsic random variability in the latent states dynamics, as well as (ii) the variability between individuals, and also (iii) account for measurement error. This flexibility gives rise to methodological and computational difficulties.
Fully Bayesian inference for nonlinear SDEMEMs is complicated by the typical intractability of the observed data likelihood which motivates the use of sampling-based approaches such as Markov chain Monte Carlo. A Gibbs sampler is proposed to target the marginal posterior of all parameters of interest. The algorithm is made computationally efficient through careful use of blocking strategies, particle filters (sequential Monte Carlo) and correlated pseudo-marginal approaches. The resulting methodology is is flexible, general and is able to deal with a large class of nonlinear SDEMEMs [1]. In a more recent work [2], we also explored ways to make inference even more scalable to an increasing number of individuals, while also dealing with state-space models driven by other stochastic dynamic models than SDEs, eg Markov jump processes and nonlinear solvers typically used in systems biology.
[1] S. Wiqvist, A. Golightly, AT McLean, U. Picchini (2020). Efficient inference for stochastic differential mixed-effects models using correlated particle pseudo-marginal algorithms, CSDA, https://doi.org/10.1016/j.csda.2020.107151
[2] S. Persson, N. Welkenhuysen, S. Shashkova, S. Wiqvist, P. Reith, G. W. Schmidt, U. Picchini, M. Cvijovic (2021). PEPSDI: Scalable and flexible inference framework for stochastic dynamic single-cell models, bioRxiv doi:10.1101/2021.07.01.450748.
Bayesian inference for mixed-effects models driven by SDEs and other stochast...Umberto Picchini
An important, and well studied, class of stochastic models is given by stochastic differential equations (SDEs). In this talk, we consider Bayesian inference based on measurements from several individuals, to provide inference at the "population level" using mixed-effects modelling. We consider the case where dynamics are expressed via SDEs or other stochastic (Markovian) models. Stochastic differential equation mixed-effects models (SDEMEMs) are flexible hierarchical models that account for (i) the intrinsic random variability in the latent states dynamics, as well as (ii) the variability between individuals, and also (iii) account for measurement error. This flexibility gives rise to methodological and computational difficulties.
Fully Bayesian inference for nonlinear SDEMEMs is complicated by the typical intractability of the observed data likelihood which motivates the use of sampling-based approaches such as Markov chain Monte Carlo. A Gibbs sampler is proposed to target the marginal posterior of all parameters of interest. The algorithm is made computationally efficient through careful use of blocking strategies, particle filters (sequential Monte Carlo) and correlated pseudo-marginal approaches. The resulting methodology is is flexible, general and is able to deal with a large class of nonlinear SDEMEMs [1]. In a more recent work [2], we also explored ways to make inference even more scalable to an increasing number of individuals, while also dealing with state-space models driven by other stochastic dynamic models than SDEs, eg Markov jump processes and nonlinear solvers typically used in systems biology.
[1] S. Wiqvist, A. Golightly, AT McLean, U. Picchini (2020). Efficient inference for stochastic differential mixed-effects models using correlated particle pseudo-marginal algorithms, CSDA, https://doi.org/10.1016/j.csda.2020.107151
[2] S. Persson, N. Welkenhuysen, S. Shashkova, S. Wiqvist, P. Reith, G. W. Schmidt, U. Picchini, M. Cvijovic (2021). PEPSDI: Scalable and flexible inference framework for stochastic dynamic single-cell models, bioRxiv doi:10.1101/2021.07.01.450748.
Presentation of Birnbaum's Likelihood Principle foundational paper at the Reading Statistical Classics seminar, Jan. 20, 2013, Université Paris-Dauphine
Data fusion is the process of combining data from different sources to enhance the utility of the combined product. In remote sensing, input data sources are typically massive, noisy, and have different spatial supports and sampling characteristics. We take an inferential approach to this data fusion problem: we seek to infer a true but not directly observed spatial (or spatio-temporal) field from heterogeneous inputs. We use a statistical model to make these inferences, but like all models it is at least somewhat uncertain. In this talk, we will discuss our experiences with the impacts of these uncertainties and some potential ways addressing them.
Scalable inference for a full multivariate stochastic volatilitySYRTO Project
Scalable inference for a full multivariate stochastic volatility
P. Dellaportas, A. Plataniotis and M. Titsias UCL(London), AUEB(Athens), AUEB(Athens)
Final SYRTO Conference - Université Paris1 Panthéon-Sorbonne
February 19, 2016
Accelerating Pseudo-Marginal MCMC using Gaussian ProcessesMatt Moores
The grouped independence Metropolis-Hastings (GIMH) and Markov chain within Metropolis (MCWM) algorithms are pseudo-marginal methods used to perform Bayesian inference in latent variable models. These methods replace intractable likelihood calculations with unbiased estimates within Markov chain Monte Carlo algorithms. The GIMH method has the posterior of interest as its limiting distribution, but suffers from poor mixing if it is too computationally intensive to obtain high-precision likelihood estimates. The MCWM algorithm has better mixing properties, but less theoretical support. In this paper we accelerate the GIMH method by using a Gaussian process (GP) approximation to the log-likelihood and train this GP using a short pilot run of the MCWM algorithm. Our new method, GP-GIMH, is illustrated on simulated data from a stochastic volatility and a gene network model. Our approach produces reasonable estimates of the univariate and bivariate posterior distributions, and the posterior correlation matrix in these examples with at least an order of magnitude improvement in computing time.
Seminar Talk: Multilevel Hybrid Split Step Implicit Tau-Leap for Stochastic R...Chiheb Ben Hammouda
In biochemically reactive systems with small copy numbers of one or more reactant molecules, the dynamics are dominated by stochastic effects. To approximate those systems, discrete state-space and stochastic simulation approaches have been shown to be more relevant than continuous state-space and deterministic ones. These stochastic models constitute the theory of Stochastic Reaction Networks (SRNs). In systems characterized by having simultaneously fast and slow timescales, existing discrete space-state stochastic path simulation methods, such as the stochastic simulation algorithm (SSA) and the explicit tau-leap (explicit-TL) method, can be very slow. In this talk, we propose a novel implicit scheme, split-step implicit tau-leap (SSI-TL), to improve numerical stability and provide efficient simulation algorithms for those systems. Furthermore, to estimate statistical quantities related to SRNs, we propose a novel hybrid Multilevel Monte Carlo (MLMC) estimator in the spirit of the work by Anderson and Higham (SIAM Multiscal Model. Simul. 10(1), 2012). This estimator uses the SSI-TL scheme at levels where the explicit-TL method is not applicable due to numerical stability issues, and then, starting from a certain interface level, it switches to the explicit scheme. We present numerical examples that illustrate the achieved gains of our proposed approach in this context.
We provide a comprehensive convergence analysis of the asymptotic preserving implicit-explicit particle-in-cell (IMEX-PIC) methods for the Vlasov–Poisson system with a strong magnetic field. This study is of utmost importance for understanding the behavior of plasmas in magnetic fusion devices such as tokamaks, where such a large magnetic field needs to be applied in order to keep the plasma particles on desired tracks.
Stochastic reaction networks (SRNs) are a particular class of continuous-time Markov chains used to model a wide range of phenomena, including biological/chemical reactions, epidemics, risk theory, queuing, and supply chain/social/multi-agents networks. In this context, we explore the efficient estimation of statistical quantities, particularly rare event probabilities, and propose two alternative importance sampling (IS) approaches [1,2] to improve the Monte Carlo (MC) estimator efficiency. The key challenge in the IS framework is to choose an appropriate change of probability measure to achieve substantial variance reduction, which often requires insights into the underlying problem. Therefore, we propose an automated approach to obtain a highly efficient path-dependent measure change based on an original connection between finding optimal IS parameters and solving a variance minimization problem via a stochastic optimal control formulation. We pursue two alternative approaches to mitigate the curse of dimensionality when solving the resulting dynamic programming problem. In the first approach [1], we propose a learning-based method to approximate the value function using a neural network, where the parameters are determined via a stochastic optimization algorithm. As an alternative, we present in [2] a dimension reduction method, based on mapping the problem to a significantly lower dimensional space via the Markovian projection (MP) idea. The output of this model reduction technique is a low dimensional SRN (potentially one dimension) that preserves the marginal distribution of the original high-dimensional SRN system. The dynamics of the projected process are obtained via a discrete $L^2$ regression. By solving a resulting projected Hamilton-Jacobi-Bellman (HJB) equation for the reduced-dimensional SRN, we get projected IS parameters, which are then mapped back to the original full-dimensional SRN system, and result in an efficient IS-MC estimator of the full-dimensional SRN. Our analysis and numerical experiments verify that both proposed IS (learning based and MP-HJB-IS) approaches substantially reduce the MC estimator’s variance, resulting in a lower computational complexity in the rare event regime than standard MC estimators. [1] Ben Hammouda, C., Ben Rached, N., and Tempone, R., and Wiechert, S. Learning-based importance sampling via stochastic optimal control for stochastic reaction net-works. Statistics and Computing 33, no. 3 (2023): 58. [2] Ben Hammouda, C., Ben Rached, N., and Tempone, R., and Wiechert, S. (2023). Automated Importance Sampling via Optimal Control for Stochastic Reaction Networks: A Markovian Projection-based Approach. To appear soon.
I am Bing Jr. I am a Signal Processing Assignment Expert at matlabassignmentexperts.com. I hold a Master's in Matlab Deakin University, Australia. I have been helping students with their assignments for the past 9 years. I solve assignments related to Signal Processing.
Visit matlabassignmentexperts.com or email info@matlabassignmentexperts.com. You can also call on +1 678 648 4277 for any assistance with Signal Processing Assignments.
Similar to Dependent processes in Bayesian Nonparametrics (20)
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...Subhajit Sahu
Abstract — Levelwise PageRank is an alternative method of PageRank computation which decomposes the input graph into a directed acyclic block-graph of strongly connected components, and processes them in topological order, one level at a time. This enables calculation for ranks in a distributed fashion without per-iteration communication, unlike the standard method where all vertices are processed in each iteration. It however comes with a precondition of the absence of dead ends in the input graph. Here, the native non-distributed performance of Levelwise PageRank was compared against Monolithic PageRank on a CPU as well as a GPU. To ensure a fair comparison, Monolithic PageRank was also performed on a graph where vertices were split by components. Results indicate that Levelwise PageRank is about as fast as Monolithic PageRank on the CPU, but quite a bit slower on the GPU. Slowdown on the GPU is likely caused by a large submission of small workloads, and expected to be non-issue when the computation is performed on massive graphs.
As Europe's leading economic powerhouse and the fourth-largest hashtag#economy globally, Germany stands at the forefront of innovation and industrial might. Renowned for its precision engineering and high-tech sectors, Germany's economic structure is heavily supported by a robust service industry, accounting for approximately 68% of its GDP. This economic clout and strategic geopolitical stance position Germany as a focal point in the global cyber threat landscape.
In the face of escalating global tensions, particularly those emanating from geopolitical disputes with nations like hashtag#Russia and hashtag#China, hashtag#Germany has witnessed a significant uptick in targeted cyber operations. Our analysis indicates a marked increase in hashtag#cyberattack sophistication aimed at critical infrastructure and key industrial sectors. These attacks range from ransomware campaigns to hashtag#AdvancedPersistentThreats (hashtag#APTs), threatening national security and business integrity.
🔑 Key findings include:
🔍 Increased frequency and complexity of cyber threats.
🔍 Escalation of state-sponsored and criminally motivated cyber operations.
🔍 Active dark web exchanges of malicious tools and tactics.
Our comprehensive report delves into these challenges, using a blend of open-source and proprietary data collection techniques. By monitoring activity on critical networks and analyzing attack patterns, our team provides a detailed overview of the threats facing German entities.
This report aims to equip stakeholders across public and private sectors with the knowledge to enhance their defensive strategies, reduce exposure to cyber risks, and reinforce Germany's resilience against cyber threats.
StarCompliance is a leading firm specializing in the recovery of stolen cryptocurrency. Our comprehensive services are designed to assist individuals and organizations in navigating the complex process of fraud reporting, investigation, and fund recovery. We combine cutting-edge technology with expert legal support to provide a robust solution for victims of crypto theft.
Our Services Include:
Reporting to Tracking Authorities:
We immediately notify all relevant centralized exchanges (CEX), decentralized exchanges (DEX), and wallet providers about the stolen cryptocurrency. This ensures that the stolen assets are flagged as scam transactions, making it impossible for the thief to use them.
Assistance with Filing Police Reports:
We guide you through the process of filing a valid police report. Our support team provides detailed instructions on which police department to contact and helps you complete the necessary paperwork within the critical 72-hour window.
Launching the Refund Process:
Our team of experienced lawyers can initiate lawsuits on your behalf and represent you in various jurisdictions around the world. They work diligently to recover your stolen funds and ensure that justice is served.
At StarCompliance, we understand the urgency and stress involved in dealing with cryptocurrency theft. Our dedicated team works quickly and efficiently to provide you with the support and expertise needed to recover your assets. Trust us to be your partner in navigating the complexities of the crypto world and safeguarding your investments.
【社内勉強会資料_Octo: An Open-Source Generalist Robot Policy】
Dependent processes in Bayesian Nonparametrics
1. Dependent processes
in Bayesian nonparametrics
Matteo Ruggiero
University of Torino and Collegio Carlo Alberto
Moncalieri, Feb 19 2016
0.0 0.2 0.4 0.6 0.8 1.0
time 1
0
0.029
0.059
0.088
2. 1. Motivation and general setting
BNP and discrete random probability measures
p = (p1, p2, . . .) frequencies in
∆∞ = p ∈ [0, 1]∞
:
i
pi = 1
p↓
= (p(1), p(2), . . .) ordered frequencies in
∞ = p ∈ [0, 1]∞
: p1 ≥ p2 ≥ · · · ≥ 0,
i
pi = 1
Assign law to p, which induces a distribution
on ∆∞, ∞
Otherwise assign to the indices unique labels
X1, X2, . . .
iid
∼ P0 continuous on X and define
the discrete measure
∞
i=1
piδXi
which induces a distribution on P(X)
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 3
3. 1. Motivation and general setting
BNP and discrete random probability measures
Approach 1:
model observations Yj directly with
p = (p1, p2, . . .) or P =
∞
i=1
piδXi
where Yj = Xi w.p. pi, and the (Xi, pi) are random
Approach 2:
use mixtures to yield more flexibility and possibly aim at continuous
distributions
f(y) =
X
f(y | x)P(dx) ⇒ f(y) =
∞
i=1
pif(y | Xi)
i.e. Yj ∼ f(y | Xi) w.p. pi and the (Xi, pi) are random
Use either approach as a base for estimation, uncertainty quantification,
forecasting, clustering, . . .
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 4
4. 1. Motivation and general setting
Motivation for dependent processes
Assumptions in classical BNP approach:
observations are excheangeable
observations depend on a fixed environment/state of the world
inference is static (fixed time)/carried out on single environment
Data may not satisfy these assumptions (e.g. prices dynamics)
Need for more general types of dependence
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 5
5. 1. Motivation and general setting
Partial exchangeability
Natural extension is partial exchangeability (de Finetti sense), e.g.
X1,1 X1,2 X1,3 · · ·
X2,1 X2,2 X2,3 · · ·
X3,1 X3,2 X3,3 · · ·
· · · · · · · · · · · ·
row-wise exchangeability (not overall): given i, Xi,j are exchangeable
Accommodates e.g. temporal structures
Collection of random probability measures, indexed by some covariate
Can be extended to an uncountable family
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 6
6. 1. Motivation and general setting
Dependent densities: discrete time
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 7
7. 1. Motivation and general setting
Dependent densities: discrete time
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 8
8. 1. Motivation and general setting
Dependent densities: continuous time
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 9
9. 1. Motivation and general setting
Dependent densities: continuous time
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 10
10. 1. Motivation and general setting
Dependent densities: continuous time
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 11
11. 1. Motivation and general setting
Modelling and inference with time-dependent processes
Temporal dependence structure
Partial exchangeability, for any t we have a distribution (possibly a mixture)
(Possibly multiple) data available at discrete time points
Model collection of random probability measures, forming
a discrete time process, or
a continuous-time process, with continuous paths or jumps
Nonparametric approach to allow for full flexibility
Analyse properties of the resulting model
Devise suitable strategies for
posterior computation
Carry out inference on desired quantities
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 12
12. 1. Motivation and general setting
General setting
X1, X2, . . .
iid
∼ P0 unique labels or locations in X
We are interested in time-dependent random probability measures of type
p(t) = (p1(t), p2(t), . . .) ∈ ∆∞
p↓
(t) = (p(1)(t), p(2)(t), . . .) ∈ ∞
P(t) =
∞
i=1
pi(t)δXi(t) ∈ P(X)
where t ≥ 0 represents time.
Discrete sample paths:
p, p↓
, P are countable collections of distributions, t ∈ N
Continous sample paths:
p, p↓
, P are (random) t-continuous functions from [0, ∞) to ∆∞, ∞ or P(X)
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 13
13. 2. Diffusive Dirichlet mixture models
Dirichlet process
The Dirichlet process [Ferguson 1973] extends the Dirichlet distribution from K
to infinitely many types
Can be defined via stick-breaking [Sethuraman 1994]
Vi
iid
∼ Beta(1, θ), pi = Vi
i−1
k=1
(1 − Vk)
0 1
p1 = V1 1 − V1
V2
p2 (1 − V1)(1 − V2)
V3
...
s.t. pi → 0 as i → ∞ and i≥1 pi = 1
Take Xi
iid
∼ P0 with P0 continuous on X.
Then P = ∞
i=1 piδXi is a Dirichlet process
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 15
14. 2. Diffusive Dirichlet mixture models
Dirichlet process
0.0 0.2 0.4 0.6 0.8 1.0
0.00
0.02
0.04
0.06
0.08
0.10
x
p
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 16
15. 2. Diffusive Dirichlet mixture models
Dependent Dirichlet process
Basic idea [MacEachern, 1999]
We aim at defining a process
P(t) =
∞
i=1
pi(t)δXi(t), t ≥ 0,
with Dirichlet process marginals
Handling both (p1(t), p2(t), . . .) and (X1(t), X2(t), . . .) can be non trivial.
Consider instead
P(t) =
∞
i=1
pi(t)δXi , t ≥ 0, Xi
iid
∼ P0
Atoms are fixed, but there are infinitely many of them
In practice, as many as you need
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 17
16. 2. Diffusive Dirichlet mixture models
Diffusive Dirichlet process
Take the Dirichlet stick-breaking weights
pi = Vi
i−1
k=1
(1 − Vk), Vi ∼iid
Beta(1, θ)
Substitute each component Vi ∈ [0, 1] with a diffusion {Vi(t)}t≥0 on [0, 1]
Then take
pi(t) = Vi(t)
i−1
k=1
(1 − Vk(t))
Each component needs to have Beta marginals, Vi(t) ∼ Beta(1, θ)
One-dimensional Wright–Fisher diffusions satisfy this
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 18
18. 2. Diffusive Dirichlet mixture models
Wright–Fisher diffusions
% of type 1 individuals (mutation rates: theta_1 = 2 , theta_2 = 8 )
Time (50K steps)
Statespace
0 2 4 6 8 10
0
1
Ergodic frequencies against Stationary Distribution Beta( 2 , 8 )
State space
0.0 0.2 0.4 0.6 0.8 1.0
0123
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 20
19. 2. Diffusive Dirichlet mixture models
Wright–Fisher diffusions
% of type 1 individuals (mutation rates: theta_1 = 8 , theta_2 = 8 )
Time (50K steps)
Statespace
0 2 4 6 8 10
0
1
Ergodic frequencies against Stationary Distribution Beta( 8 , 8 )
State space
0.0 0.2 0.4 0.6 0.8 1.0
0.01.53.0
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 21
20. 2. Diffusive Dirichlet mixture models
Wright–Fisher diffusions
% of type 1 individuals (mutation rates: theta_1 = 0.4 , theta_2 = 0.4 )
Time (50K steps)
Statespace
0 2 4 6 8 10
0
1
Ergodic frequencies against Stationary Distribution Beta( 0.4 , 0.4 )
State space
0.0 0.2 0.4 0.6 0.8 1.0
048
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 22
21. 2. Diffusive Dirichlet mixture models
Diffusive Dirichlet process [Mena and R. 2016]
The resulting object
P(t) =
∞
i=1
Vi(t)
i−1
k=1
(1 − Vk(t))
pi(t)
δXi , Vi(t) ∼ WF(a, b)
has Dirichlet marginals for (a, b) = (1, θ), i.e. P(t) is a DP for all t
has GEM marginals for (a, b) ∈ R2
+
has diffusive behaviour, P(t) is t-continuous in total variation
See also
Gutierrez, Mena and & R. 2016 (version with jumps)
Mena, R. & Walker 2011 (geometric weights, different marginals)
for related models
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 23
23. 2. Diffusive Dirichlet mixture models
Estimation
At each time ti we have observations (yi,1, . . . , yi,ni ).
Set up the hierarchical mixture
{Pt, t ≥ 0} ∼ diff-DP or GSB
xti | Pti ∼ Pti
yi,j | ti, xti
iid
∼ f(· | xti )
equivalently yi is drawn from the time-dependent nonparametric mixture model
fti (y) =
X
f(y|x)Pti (dx) =
∞
i=1
pti f(y | xi)
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 25
24. 2. Diffusive Dirichlet mixture models
Simulated data
True model
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 26
25. 2. Diffusive Dirichlet mixture models
Simulated data
Single data points
0 2 4 6 8 10
−202468
True model (heat map), posterior mode (solid), 95% credible intervals for the mean (dashed), 95%
quantiles of posterior density estimate (dotted).
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 27
26. 2. Diffusive Dirichlet mixture models
Simulated data
Multiple data points
0 2 4 6 8 10
−202468
True model (heat map), posterior mode (solid), 95% credible intervals for the mean (dashed), 95%
quantiles of posterior density estimate (dotted).
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 28
27. 2. Diffusive Dirichlet mixture models
Real data: S&P 500 (03/08 - 02/09)
Dependent density estimate
Heat map of estimated density (red), and mean estimate (solid)
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 29
29. 2. Diffusive Dirichlet mixture models
Real data: S&P 500 (03/08 - 02/09)
Dependent density estimate
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 31
30. populations
A different view: modelling evolving populations
A sample path of p↓
(t) = (p(1), . . . , p(7))
Time
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Frequency
Dynamic frenquencies of 7 species
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 33
31. populations
A different view: modelling evolving populations
Distinct values X1, X2, . . . are interpreted as
allelic types in genetics
plant or animal species
unique identifiers of some evolving groups
Large population → species abundances approximate diffusive behaviours
If cannot provide an a priori upper bound, assume infinitely many species
Two different approaches:
constructing stochastic models for pseudo-realistic evolutionary mechanisms
(mutation, selection, recombination, migration, . . . )
studying the association between certain
distributions and connected dynamics
Dynamics in figure are related to
a Dirichlet distribution
Can we extend them? To what extent?
With what interpretation?
Time
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Frequency
Dynamic frenquencies of 7 species
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 34
33. populations
Poisson-Dirichlet case
No. species Markov chain
(N individuals)
K
Wright-Fisher(N, K, θ)
Fisher (1930), Wright (1931)
Diffusion
(∞ individuals)
d
N → ∞
Wright-Fisher(K, θ)
Sato (1976)
stationary
w.r.t.
Dir θ
K , . . . , θ
K
Random measure
(t fixed)
∞ IMNA(θ)
Ethier and Kurtz (1981)
d K → ∞
PD(θ)
Kingman (1975)
d K → ∞
stationary
w.r.t.
Moran(N, θ)
Watterson (1976)
d
N → ∞
“
d
−→” = convergence in distribution
IMNA = infinitely many neutral alleles
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 36
34. populations
Two-parameter Poisson-Dirichlet case
No. species
∞ PD(θ, α)
Pitman (1995)
Random measure
(t fixed)
Diffusion
(∞ individuals)
IMNA(θ, α)
Petrov (2009)
stationary
w.r.t.
?? Moran(N, θ, α)
R. and Walker (2009)
d
N → ∞
Markov chain
(N individuals)
?? WF(K, θ, α)
Costantini, De Blasi,
Ethier, R., Span`o (2016)
d K → ∞
K ?? WF(N, K, θ, α)
Costantini, De Blasi,
Ethier, R., Span`o (2016)
d
N → ∞
stationary
w.r.t. ??
d K → ∞
Remarks:
IMNA = infinitely many neutral allelesBased on Pitman’s generalized P´olya urn schemeMutation and immigration
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 37
38. The propagation mixture
Prior X ∼ πα := Gamma(α1, α2)
Likelihood Y | X ∼ Poisson(X)
Posterior X | Y1, . . . , Yn ∼ πα,n := Gamma α1 +
n
i=1
yi, α2 + n
Propagation mixture [Papaspiliopoulos & R. 2014]
ψt(πα,n) := πα,n(x)Pt(x, dx )
is given by
ψt(πα,n) =
n
j=0
pt(n, j)Gamma α1 +
n
i=0
yi − j, α2 + n − st
for appropriate time-varying weights pt(n, j)
Can be extended to infinite dimensional models [Papaspiliopoulos, R. & Span`o
2016]
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 42
41. Some references
Costantini, De Blasi, Ethier, R. and Span`o (2016).
Wright–Fisher construction of the two-parameter Poisson–Dirichlet diffusion.
arXiv:1601.06064
Gutierrez, Mena & R. (2016).
A time dependent Bayesian nonparametric model for air quality analysis.
Comput. Statist. Data Anal.
Mena & R. (2016).
Dynamic density estimation with diffusive Dirichlet mixtures. Bernoulli
Mena, R. & Walker (2011).
Geometric stick-breaking processes for continuous-time Bayesian nonparametric modeling.
J. Statist. Plann. Inf.
Papaspiliopoulos & R. (2014).
Optimal filtering and the dual process. Bernoulli
Papaspiliopoulos, R. & Span`o (2014).
Filtering hidden Markov measures. arXiv:1411.4944
R. & Walker (2009).
Countable representation for infinite dimensional diffusions derived from the
two-parameter Poisson–Dirichlet process. Electr. Comm. Probab.
For more info: www.matteoruggiero.it
Matteo Ruggiero (Unito & CCA) Dependent processes in BNP 45