This document summarizes the Ph.D. work of Apostolos Chalkis on efficient geometric random walks for high-dimensional sampling from convex bodies. The goals are to develop algorithms and implementations of random walks like reflective Hamiltonian Monte Carlo and the billiard walk to sample from log-concave distributions restricted to polytopes, zonotopes, and spectrahedra. The work has applications in areas like computational geometry, computational finance, and systems biology. Open-source software packages in C++, R, and Python have been created to implement the sampling algorithms.
Sampling Spectrahedra: Volume Approximation and OptimizationApostolos Chalkis
My talk to SIAM Conference on Applied Algebraic Geometry (AG21) on volume approximation of spectrahedra and convex optimization with randomized methods based on MCMC sampling with geometric random walks
Practical volume estimation of polytopes by billiard trajectories and a new a...Apostolos Chalkis
A new randomized method to approximate the volume of a convex polytope based on simulated annealing for cooling convex bodies and MCMC sampling with geometric random walks
Bagging Exponential Smoothing procedures have recently arisen as an innovative way to improve forecast accuracy. The idea is to use Bootstrap to generate multiple versions of the time series and, subsequently, apply an Exponential Smoothing (ETS) method to produce forecasts for each of them. The final result is obtained aggregating the forecasts. The main drawback of existing procedures is that Bagging itself does not avoid generating highly correlated ensembles that might affect the forecast error. In this paper we propose and evaluate procedures that try to enhance existing Bagging Exponential Smoothing methods by an addition of a clustering phase. The general idea is to generate Bootstrapped versions of the series and use clusters to select series that are less similar among each other. The expectation is that this would reduce the covariance and, consequently, the forecast error. Since there are several cluster algorithms and dissimilarity measures, we consider some of them in the study. The proposed procedures were evaluated on monthly, quarterly and yearly data from the M3-competition. The results were quite promising, indicating that the introduction of a cluster phase in the Bagging Exponential Smoothing procedures can reduce the forecast error.
Sampling Spectrahedra: Volume Approximation and OptimizationApostolos Chalkis
My talk to SIAM Conference on Applied Algebraic Geometry (AG21) on volume approximation of spectrahedra and convex optimization with randomized methods based on MCMC sampling with geometric random walks
Practical volume estimation of polytopes by billiard trajectories and a new a...Apostolos Chalkis
A new randomized method to approximate the volume of a convex polytope based on simulated annealing for cooling convex bodies and MCMC sampling with geometric random walks
Bagging Exponential Smoothing procedures have recently arisen as an innovative way to improve forecast accuracy. The idea is to use Bootstrap to generate multiple versions of the time series and, subsequently, apply an Exponential Smoothing (ETS) method to produce forecasts for each of them. The final result is obtained aggregating the forecasts. The main drawback of existing procedures is that Bagging itself does not avoid generating highly correlated ensembles that might affect the forecast error. In this paper we propose and evaluate procedures that try to enhance existing Bagging Exponential Smoothing methods by an addition of a clustering phase. The general idea is to generate Bootstrapped versions of the series and use clusters to select series that are less similar among each other. The expectation is that this would reduce the covariance and, consequently, the forecast error. Since there are several cluster algorithms and dissimilarity measures, we consider some of them in the study. The proposed procedures were evaluated on monthly, quarterly and yearly data from the M3-competition. The results were quite promising, indicating that the introduction of a cluster phase in the Bagging Exponential Smoothing procedures can reduce the forecast error.
A new practical algorithm for volume estimation using annealing of convex bodiesVissarion Fisikopoulos
We study the problem of estimating the volume of convex polytopes, focusing on H- and V-polytopes, as well as zonotopes. Although a lot of effort is devoted to practical algorithms for H-polytopes there is no such method for the latter two representations. We propose a new, practical algorithm for all representations, which is faster than existing methods. It relies on Hit-and-Run sampling, and combines a new simulated annealing method with the Multiphase Monte Carlo (MMC) approach. Our method introduces the following key features to make it adaptive: (a) It defines a sequence of convex bodies in MMC by introducing a new annealing schedule, whose length is shorter than in previous methods with high probability, and the need of computing an enclosing and an inscribed ball is removed; (b) It exploits statistical properties in rejection-sampling and proposes a better empirical convergence criterion for specifying each step; (c) For zonotopes, it may use a sequence of convex bodies for MMC different than balls, where the chosen body adapts to the input. We offer an open-source, optimized C++ implementation, and analyze its performance to show that it outperforms state-of-the-art software for H-polytopes by Cousins-Vempala (2016) and Emiris-Fisikopoulos (2018), while it undertakes volume computations that were intractable until now, as it is the first polynomial-time, practical method for V-polytopes and zonotopes that scales to high dimensions (currently 100). We further focus on zonotopes, and characterize them by their order (number of generators over dimension), because this largely determines sampling complexity. We analyze a related application, where we evaluate methods of zonotope approximation in engineering.
In general dimension, there is no known total polynomial algorithm for either convex hull or vertex enumeration, i.e. an algorithm whose complexity depends polynomially on the input and output sizes. It is thus important to identify problems (and polytope representations) for which total polynomial-time algorithms can be obtained. We offer the first total polynomial-time algorithm for computing the edge-skeleton (including vertex enumeration) of a polytope given by an optimization or separation oracle, where we are also given a superset of its edge directions. We also offer a space-efficient variant of our algorithm by employing reverse search. All complexity bounds refer to the (oracle) Turing machine model. There is a number of polytope classes naturally defined by oracles; for some of them neither vertex nor facet representation is obvious. We consider two main applications, where we obtain (weakly) total polynomial-time algorithms: Signed Minkowski sums of convex polytopes, where polytopes can be subtracted provided the signed sum is a convex polytope, and computation of secondary, resultant, and discriminant polytopes. Further applications include convex combinatorial optimization and convex integer programming, where we offer a new approach, thus removing the complexity's exponential dependence in the dimension.
Computing the volume of a convex body is a fundamental problem in computational geometry and optimization. In this talk we discuss the computational complexity of this problem from a theoretical as well as practical point of view. We show examples of how volume computation appear in applications ranging from combinatorics to algebraic geometry.
Next, we design the first practical algorithm for polytope volume approximation in high dimensions (few hundreds).
The algorithm utilizes uniform sampling from a convex region and efficient boundary polytope oracles.
Interestingly, our software provides a framework for exploring theoretical advances since it is believed, and our experiments provide evidence for this belief, that the current asymptotic bounds are unrealistically high.
In this talk, I address two new ideas in sampling geometric objects. The first is a new take on adaptive sampling with respect to the local feature size, i.e., the distance to the medial axis. We recently proved that such samples acn be viewed as uniform samples with respect to an alternative metric on the Euclidean space. The second is a generalization of Voronoi refinement sampling. There, one also achieves an adaptive sample while simultaneously "discovering" the underlying sizing function. We show how to construct such samples that are spaced uniformly with respect to the $k$th nearest neighbor distance function.
Sensors and Samples: A Homological ApproachDon Sheehy
In their seminal work on homological sensor networks, de Silva and Ghrist showed the surprising fact that its possible to certify the coverage of a coordinate free sensor network even with very minimal knowledge of the space to be covered. We give a new, simpler proof of the de Silva-Ghrist Topological Coverage Criterion that eliminates any assumptions about the smoothness of the boundary of the underlying space, allowing the results to be applied to much more general problems. The new proof factors the geometric, topological, and combinatorial aspects of this approach. This factoring reveals an interesting new connection between the topological coverage condition and the notion of weak feature size in geometric sampling theory. We then apply this connection to the problem of showing that for a given scale, if one knows the number of connected components and the distance to the boundary, one can also infer the higher betti numbers or provide strong evidence that more samples are needed. This is in contrast to previous work which merely assumed a good sample and gives no guarantees if the sampling condition is not met.
A new practical algorithm for volume estimation using annealing of convex bodiesVissarion Fisikopoulos
We study the problem of estimating the volume of convex polytopes, focusing on H- and V-polytopes, as well as zonotopes. Although a lot of effort is devoted to practical algorithms for H-polytopes there is no such method for the latter two representations. We propose a new, practical algorithm for all representations, which is faster than existing methods. It relies on Hit-and-Run sampling, and combines a new simulated annealing method with the Multiphase Monte Carlo (MMC) approach. Our method introduces the following key features to make it adaptive: (a) It defines a sequence of convex bodies in MMC by introducing a new annealing schedule, whose length is shorter than in previous methods with high probability, and the need of computing an enclosing and an inscribed ball is removed; (b) It exploits statistical properties in rejection-sampling and proposes a better empirical convergence criterion for specifying each step; (c) For zonotopes, it may use a sequence of convex bodies for MMC different than balls, where the chosen body adapts to the input. We offer an open-source, optimized C++ implementation, and analyze its performance to show that it outperforms state-of-the-art software for H-polytopes by Cousins-Vempala (2016) and Emiris-Fisikopoulos (2018), while it undertakes volume computations that were intractable until now, as it is the first polynomial-time, practical method for V-polytopes and zonotopes that scales to high dimensions (currently 100). We further focus on zonotopes, and characterize them by their order (number of generators over dimension), because this largely determines sampling complexity. We analyze a related application, where we evaluate methods of zonotope approximation in engineering.
In general dimension, there is no known total polynomial algorithm for either convex hull or vertex enumeration, i.e. an algorithm whose complexity depends polynomially on the input and output sizes. It is thus important to identify problems (and polytope representations) for which total polynomial-time algorithms can be obtained. We offer the first total polynomial-time algorithm for computing the edge-skeleton (including vertex enumeration) of a polytope given by an optimization or separation oracle, where we are also given a superset of its edge directions. We also offer a space-efficient variant of our algorithm by employing reverse search. All complexity bounds refer to the (oracle) Turing machine model. There is a number of polytope classes naturally defined by oracles; for some of them neither vertex nor facet representation is obvious. We consider two main applications, where we obtain (weakly) total polynomial-time algorithms: Signed Minkowski sums of convex polytopes, where polytopes can be subtracted provided the signed sum is a convex polytope, and computation of secondary, resultant, and discriminant polytopes. Further applications include convex combinatorial optimization and convex integer programming, where we offer a new approach, thus removing the complexity's exponential dependence in the dimension.
Computing the volume of a convex body is a fundamental problem in computational geometry and optimization. In this talk we discuss the computational complexity of this problem from a theoretical as well as practical point of view. We show examples of how volume computation appear in applications ranging from combinatorics to algebraic geometry.
Next, we design the first practical algorithm for polytope volume approximation in high dimensions (few hundreds).
The algorithm utilizes uniform sampling from a convex region and efficient boundary polytope oracles.
Interestingly, our software provides a framework for exploring theoretical advances since it is believed, and our experiments provide evidence for this belief, that the current asymptotic bounds are unrealistically high.
In this talk, I address two new ideas in sampling geometric objects. The first is a new take on adaptive sampling with respect to the local feature size, i.e., the distance to the medial axis. We recently proved that such samples acn be viewed as uniform samples with respect to an alternative metric on the Euclidean space. The second is a generalization of Voronoi refinement sampling. There, one also achieves an adaptive sample while simultaneously "discovering" the underlying sizing function. We show how to construct such samples that are spaced uniformly with respect to the $k$th nearest neighbor distance function.
Sensors and Samples: A Homological ApproachDon Sheehy
In their seminal work on homological sensor networks, de Silva and Ghrist showed the surprising fact that its possible to certify the coverage of a coordinate free sensor network even with very minimal knowledge of the space to be covered. We give a new, simpler proof of the de Silva-Ghrist Topological Coverage Criterion that eliminates any assumptions about the smoothness of the boundary of the underlying space, allowing the results to be applied to much more general problems. The new proof factors the geometric, topological, and combinatorial aspects of this approach. This factoring reveals an interesting new connection between the topological coverage condition and the notion of weak feature size in geometric sampling theory. We then apply this connection to the problem of showing that for a given scale, if one knows the number of connected components and the distance to the boundary, one can also infer the higher betti numbers or provide strong evidence that more samples are needed. This is in contrast to previous work which merely assumed a good sample and gives no guarantees if the sampling condition is not met.
Cusps of the Kähler moduli space and stability conditions on K3 surfacesHeinrich Hartmann
Presentation about the paper with the same title http://arxiv.org/abs/1012.3121
Abstract:
In [Ma1] S. Ma established a bijection between Fourier--Mukai partners of a K3 surface and cusps of the K\"ahler moduli space. The K\"ahler moduli space can be described as a quotient of Bridgeland's stability manifold. We study the relation between stability conditions σ near to a cusp and the associated Fourier--Mukai partner Y in the following ways. (1) We compare the heart of σ to the heart of coherent sheaves on Y. (2) We construct Y as moduli space of σ-stable objects.
An appendix is devoted to the group of auto-equivalences of the derived category which respect the component Stab†(X) of the stability manifold
Approximation Algorithms for the Directed k-Tour and k-Stroll ProblemsSunny Kr
In the Asymmetric Traveling Salesman Problem (ATSP), the input is a directed n-vertex graph G = (V; E) with nonnegative edge lengths, and the goal is to nd a minimum-length tour, visiting
each vertex at least once. ATSP, along with its undirected counterpart, the Traveling Salesman
problem, is a classical combinatorial optimization problem
In the study of probabilistic integrators for deterministic ordinary differential equations, one goal is to establish the convergence (in an appropriate topology) of the random solutions to the true deterministic solution of an initial value problem defined by some operator. The challenge is to identify the right conditions on the additive noise with which one constructs the probabilistic integrator, so that the convergence of the random solutions has the same order as the underlying deterministic integrator. In the context of ordinary differential equations, Conrad et. al. (Stat.
Comput., 2017), established the mean square convergence of the solutions for globally Lipschitz vector fields, under the assumptions of i.i.d., state-independent, mean-zero Gaussian noise. We extend their analysis by considering vector fields that need not be globally Lipschitz, and by
considering non-Gaussian, non-i.i.d. noise that can depend on the state and that can have nonzero mean. A key assumption is a uniform moment bound condition on the noise. We obtain convergence in the stronger topology of the uniform norm, and establish results that connect this topology to the regularity of the additive noise. Joint work with A. M. Stuart (Caltech), T. J. Sullivan (Free University of Berlin).
Welcome to TechSoup New Member Orientation and Q&A (May 2024).pdfTechSoup
In this webinar you will learn how your organization can access TechSoup's wide variety of product discount and donation programs. From hardware to software, we'll give you a tour of the tools available to help your nonprofit with productivity, collaboration, financial management, donor tracking, security, and more.
Francesca Gottschalk - How can education support child empowerment.pptxEduSkills OECD
Francesca Gottschalk from the OECD’s Centre for Educational Research and Innovation presents at the Ask an Expert Webinar: How can education support child empowerment?
Introduction to AI for Nonprofits with Tapp NetworkTechSoup
Dive into the world of AI! Experts Jon Hill and Tareq Monaur will guide you through AI's role in enhancing nonprofit websites and basic marketing strategies, making it easy to understand and apply.
Macroeconomics- Movie Location
This will be used as part of your Personal Professional Portfolio once graded.
Objective:
Prepare a presentation or a paper using research, basic comparative analysis, data organization and application of economic information. You will make an informed assessment of an economic climate outside of the United States to accomplish an entertainment industry objective.
Read| The latest issue of The Challenger is here! We are thrilled to announce that our school paper has qualified for the NATIONAL SCHOOLS PRESS CONFERENCE (NSPC) 2024. Thank you for your unwavering support and trust. Dive into the stories that made us stand out!
How to Make a Field invisible in Odoo 17Celine George
It is possible to hide or invisible some fields in odoo. Commonly using “invisible” attribute in the field definition to invisible the fields. This slide will show how to make a field invisible in odoo 17.
1. Efficient geometric random walks
for high-dimensional sampling from convex bodies
Apostolos Chalkis
Ph.D. supervisor: Prof. Ioannis Z. Emiris
National & Kapodistrian U. Athens, Greece
December 20, 2021
2. Problem: Truncated log-concave sampling
Definition. Let π(x) ∝ e−f (x), where f : Rd → R is a convex
function. π(x) is called log-concave (LC) probability density.
We consider the case where π(x) is restricted in a
convex body K ⊂ Rd .
Important examples: Uniform, Gaussian, Boltzmann.
3. Why is sampling an important problem?
Several problems can be reduced to sampling:
Convex optimization [Ma,Chen,Flammarion,Jordan’19]
Volume approximation [M. Dyer, A. Frieze, and R. Kannan’91]
Multivariate integration [Lovasz,Vempala’06]
Bayesian inference [Gelman, Carlin, Stern, Dunson, Rubin’95]
Sampling also appears in plenty of applications
4. Why geometric random walks?
Alternative methods:
Acceptance rejection sampling [Casella,Christian,Wells’04]
Adaptive Rejection sampling [Gilks,Wild’92]
Slice sampling [Neal’03]
Inverse transform sampling [Olver,Townsend’13]
typically, do not scale beyond dimension d ≥ 20.
5. Geometric Random Walks
A Geometric Random Walk starts at some interior point and at
each step moves to a ”neighboring” point, chosen according to
some distribution depending only on the current point.
Implementation of Billiard Walk
step.
Uniform sampling via the
Billiard Walk.
6. Goals of my Ph.D.
Problem: Sample from a LC distribution restricted to a
convex body
Algorithmic and complexity results.
Efficient implementations of geometric random walks.
Practical, randomized algorithms to address challenging
problems in:
– Computational geometry
– Computational Finance
– Systems biology
7. Summary of results
Support and develop random walks to sample from
log-concave distributions restricted in:
H-polytopes
V-polytopes
Zonotopes (Z-polytopes)
Spectrahedra
Applications of log-concave sampling:
Volume approximation of convex bodies
Systems biology (metabolic network analysis)
Finance (crisis detection, portfolio optimization/scoring)
Convex optimization
Open-source software:
C++/R package volesti
github.com/GeomScale/volesti
python package dingo
github.com/GeomScale/dingo
9. Open-source internships
Google Summer of Code:
State-of-the-art geometric random walks for sampling from
high dimensional bodies in R, 2018.
State-of-the-art algorithms in R for volume computation and
sampling in high dimensions, 2019.
Tweag - Software Innovation Lab:
Develop dingo, a Python package to analyze metabolic
networks, 2021.
10. GeomScale org
– GeomScale/volesti
volume approximation & sampling
from convex bodies
[C,Fisikopoulos’21]
– GeomScale/dingo
analyze metabolic networks with
MCMC sampling
Co-founders: V. Fisikopoulos & E. Tsigaridas
————————————————————————————
One of the 200 mentoring organizations supported by
iii Google Summer of Code (GSoC) 2020 & 2021.
Mentor 8 GSoC projects.
NumFOCUS Affiliated Project.
Support an open community (>130 members).
More than 15 000 lines of code.
13. Convex polytope representations
V-polytope: the convex hull of a set of points in Rd .
Applications:
Systems biology (metabolic networks) [Caso,Montañez’13]
Biogeography
[Barnagaud,Kissling,Tsirogiannis,Fisikopoulos,Villéger,Sekercioglu,Svenning’17]
14. Convex polytope representations
Z-polytope: Minkowski sum of k d-dimensional segments.
A zonotope is a centrally symmetric convex body.
Applications:
Autonomous driving [Althoff,Dolan’14]
Human-robot collaboration [Pereira,Althof’15]
Neural networks [Anderson,Pailoor,Dillig,Chaudhuri’19]
15. Spectrahedron
A spectrahedron S ⊂ Rd is the feasible set of a linear matrix
inequality. If Ai are symmetric matrices in Rm×m and
F(x) = A0 + x1A1 + · · · + xd Ad ,
then S = {x ∈ Rd | F(x) 0}.
A 3D elliptope
S is the feasible set of a Semidefinite Program (SDP)
16. Random walks for
truncated log-concave sampling
Year Authors Random walk Mixing time∗
Distribution
[Smith’86] Hit-and-Run O∗
(d3
) any LC
[Berbee,Smith’87] Coordinate Hit-and-Run O∗
(d10
) any LC
[Lovasz,Simonovits’90] Ball walk O∗
(d3
) any LC
[Kannan,Narayanan’12] Dikin walk O∗
(d2
) uniform (H-polytope)
[Polyak,Dabbene’14] Billiard walk ?? uniform
[Afshar,Domke’15] Reflective HMC ?? any LC (polytopes)
[Lee,Vempala’16] Geodesic walk O(md3/4
) uniform (H-polytope)
[Lee,Vempala’17] Remannian HMC O∗
(md2/3
) uniform (H-polytopes)
[Chen,Dwivedi,Wainwright,Yu’19] John walk O∗
(d5/2
) uniform (H-polytope)
[Chen,Dwivedi,Wainwright,Yu’19] Vaidya walk O(m1/2
d3/2
) uniform (H-polytope)
Table 1
Cost per sample: cost per step × mixing time (#steps).
The cost per step depends on the convex body.
Hit-and-Run (HR): widely used well studied.
Coordinate Hit-and-Run (CDHR): has been proven more
efficient than HR in practice.
Existing software uses either CDHR or HR (H-polytopes).
∗
total variation mixing time
17. Contributions
For Reflective Hamiltonian Monte Carlo (ReHMC):
We prove that it converges to the target distribution when
the latter is truncated by a spectrahedron.
For both (ReHMC) and Billiard walk:
We improve the cost per step for H-polytopes.
We provide efficient operations for V-, Z-polytopes and
spectrahedra.
We offer efficient parameterizations.
We experimentally show that they outperform all the random
walks in Table 1.
18. Hit-and-Run(K, x, π): Convex body K ⊂ Rd , point x ∈ K,
PDF π : Rd → R+
1 Pick uniformly a line ` through x.
2 return a random point on the chord ` ∩ K chosen from
the distribution π`,π restricted in ` ∩ K.
P
x
B
`
19. Hit-and-Run(K, x, π): Convex body K ⊂ Rd , point x ∈ K,
PDF π : Rd → R+
1 Pick uniformly a line ` through x.
2 return a random point on the chord ` ∩ K chosen from
the distribution π`,π restricted in ` ∩ K.
P
`
x
20. Hit-and-Run(K, x, π): Convex body K ⊂ Rd , point x ∈ K,
PDF π : Rd → R+
1 Pick uniformly a line ` through x.
2 return a random point on the chord ` ∩ K chosen from
the distribution π`,π restricted in ` ∩ K.
P
`
x
21. Coordinate Hit-and-Run(K, x, π): Convex body K ⊂ Rd ,
point x ∈ K, PDF π : Rd → R+
1 Pick a uniformly random axis direction ei .
2 Consider the line ` through x with direction ei .
3 return a random point on the chord ` ∩ K chosen from
the distribution π`,π restricted in ` ∩ K.
`
p q
22. Hamiltonian Monte Carlo
Being at p ∈ K, HMC introduces an auxiliary random variable
v ∈ Rd and generates samples from the joint density
π(p, v) = π(v|p)π(p),
Marginalize out v, then recover the target dist. π(p).
Consider v ∼ N(0, Id );
PDF π(p, v) = e−H(p,v) defines a Hamiltonian,
H(p, v) = − log π(p, v) = − log π(p) +
1
2
|v|2
,
23. Hamiltonian Monte Carlo
HMC simulates a particle moving in a conservative field
determined by − log π(p) and −∇ log π(p).
HMC, starting from a position p, generates a new state:
1 Draw a value for the momentum, v ∼ N(0, Id )
2 (p, v) is given by the Hamilton’s system of ODE:
dp
dt
=
∂H(p, v)
∂v
dv
dt
= −
∂H(p, v)
∂p
⇒
dp(t)
dt = v(t)
dv(t)
dt = −∇ log π(p)
(1)
Solve (1) with
–Euler methods (e.g., Leapfrog) [Neal’12] or,
–Collocation method [Vempala,Lee,Song’18].
24. Reflective Hamiltonian Monte Carlo (ReHMC)
When the density is restricted in a convex body K then HMC
trajectory stays inside K by using boundary reflections.
Case of Leapfrog method
π(x) Discrete Hamiltonian
trajectory
We pre-select the number of Leapfrog steps
[C,Fisikopoulos,Papachristou,Tsigaridas’21]
25. Reflective Hamiltonian Monte Carlo (ReHMC)
When the density is restricted in a convex body K then HMC
trajectory stays inside K by using boundary reflections.
Case of collocation method
π(x)
Polynomial Hamiltonian
trajectory
We randomly select the integration time in each steps
26. Billiard walk - Uniform case
BW(P, x): Convex body K ⊂ Rd
, point x ∈ K
1 Generate the length of the trajectory L ∼ D.
2 Pick a uniform direction v to define the trajectory.
3 The trajectory reflects on the boundary if necessary.
4 return the end of the trajectory as pi+1.
27. Billiard walk - Uniform case
BW(P, x): Convex body K ⊂ Rd
, point x ∈ K
1 Generate the length of the trajectory L ∼ D.
2 Pick a uniform direction v to define the trajectory.
3 The trajectory reflects on the boundary if necessary.
4 return the end of the trajectory as pi+1.
28. Billiard walk - Uniform case
BW(P, x): Convex body K ⊂ Rd
, point x ∈ K
1 Generate the length of the trajectory L ∼ D.
2 Pick a uniform direction v to define the trajectory.
3 The trajectory reflects on the boundary if necessary.
4 return the end of the trajectory as pi+1.
29. Billiard walk - Uniform case
BW(P, x): Convex body K ⊂ Rd
, point x ∈ K
1 Generate the length of the trajectory L ∼ D.
2 Pick a uniform direction v to define the trajectory.
3 The trajectory reflects on the boundary if necessary.
4 return the end of the trajectory as pi+1.
30. Billiard walk - Uniform case
BW(P, x): Convex body K ⊂ Rd
, point x ∈ K
1 Generate the length of the trajectory L ∼ D.
2 Pick a uniform direction v to define the trajectory.
3 The trajectory reflects on the boundary if necessary.
4 return the end of the trajectory as pi+1.
31. Billiard walk - Uniform case
BW(P, x): Convex body K ⊂ Rd
, point x ∈ K
1 Generate the length of the trajectory L ∼ D.
2 Pick a uniform direction v to define the trajectory.
3 The trajectory reflects on the boundary if necessary.
4 return the end of the trajectory as pi+1.
32. Geometric and algebraic oracles
The implementation of a geometric random walk requires:
Membership oracle
Boundary (intersection) oracle
Reflection oracle
34. Membership oracle
membership(F, p): An LMI F(x) 0 ⇔ A0+x1A1+· · ·+xd Ad 0
representing a spectrahedron S and a point p ∈ Rd
.
1 λmin ← smallest eigenvalue of F(p).
2 if λmin ≥ 0 return true else return false.
35. Boundary oracle
intersection(F, Φ(t)): An LMI F(x) 0 ⇔ A0 + x1A1 + · · · +
xd Ad 0 representing a spectrahedron S,
Φ : t 7→ Φ(t) := (p1(t), . . . , pd (t)) parameterization of a polynomial
curve, where pi (t) =
Pni
j=0 pi,j tj
, and Φ(0) ∈ S.
1 Solve the polynomial eigenvalue problem
F(Φ(t)) x = 0 ⇔ (B0 + tB1 + · · · + td
Bd )x = 0,
where Bk =
Pd
j=1 pj,k Aj
2 Smallest positive and largest negative eigenvalues λ−
max , λ+
min
3 return the boundary points F(Φ(λ−
max )) and F(Φ(λ+
min))
36. Reflection oracle
reflection(F, Φ(t), λ+): An LMI F(x) 0 ⇔ A0 + x1A1 + · · · +
xd Ad 0 representing spectrahedron S,
Φ(t) parameterization of a polynomial curve,
λ+ s.t. Φ(λ+) ∈ ∂S
1 Let the boundary point p+ = Φ(λ+)
2 Let w = ∇ det(F(p+)) = c · (s
A1s, · · · , s
Ad s),
s vector in the kernel of F(p+)
3 return the direction of the reflection
s+ ← dΦ
dt (t+) − 2 h∇dΦ
dt (t+), wi w
37. Per-step complexity
Random walk per-step Complexity
HR O(mω + m log(1/) + dm2)
Coordinate HR O(mω + m log(1/) + m2)
Billiard walk e
O(ρ(mω + m log(1/) + dm2))
ReHMC (collocation) e
O(ρ((nm)ω + mn log(1/) + dnm2))
ReHMC (leapfrog) e
O(Lρ(mω + m log(1/) + dm2))
[C,Emiris,Fisikopoulos,Repouskos,Tsigaridas’20]
m: size of the matrices Ai in LMI
d: dimension
n: degree of the polynomial curve
ρ: number of reflections
: accuracy to approximate the intersection with the boundary
ω: exponent in the complexity of matrix multiplication
L: number of leapfrog steps
38. Optimization - Semidefinite Programs
In an SDP we minimize a linear function of a variable x ∈ Rd
subject to an Linear Matrix Inequality (LMI):
min cT
x
subject to F(x) = A0 + x1A1 + · · · + xnAn 0
The feasible set is a spectrahedron S = {x ∈ Rd | F(x) 0}.
Semidefinite programming generalizes Linear programming.
39. Optimization via Exponential sampling
Problem: Minimize a linear function f (x) = cT x in
spectrahedron S.
Answer: Sample from π(x) ∝ e−cT x/T restricted in S, for
T = T0 · · · TM.
T0 T1 T2 T3
Task: Compute a sequence of Ti ∈ R+ of length M s.t. a
sample from πTM
is close to the optimal solution
with high probability.
40. Simulated Annealing
Convergence to the optimal solution
πi (x) ∝ e−cT x/Ti
Starting with T0 = R, where S ⊂ RBd (uniform distribution).
Ti = Ti−1(1 − 1
√
d
), i ∈ [M] (Ti−1 is a warm start for Ti ).
M = O∗(
√
d) phases to obtain a solution |fM − f ∗| ≤
Only Hit-and-Run has been used in previous work
[Kalai,Vempala’06].
42. Faster scheduling
In practice, we speedup the temperature schedule by setting,
Ti = T0
1 −
1
dk
i
, i ∈ [I]. (2)
In theory, k = 1/2 (Hit-and-Run).
Experimentally, for ReHMC, Ti results to a warm starting
point for Ti+1, when k 1/2.
43. sdpa Vs volesti
Random generator of Spectrahedra [Polyak,Dabbene’14].
Experimental branch in GeomScale/volesti [C,Fisikopoulos,Tsigaridas]
45. In our cells...
We call both the inputs (reactants) and the outputs (products)
of a chemical reaction, metabolites.
In every cell of our body
thousands of chemical reactions are taking place!
46. The reactions interact
A small fragment of the human metabolic network
Q: How can we model all the interactions between chemical
sdf reactions in an organism?
A: Computational geometry can help!
47. Key Concept of a metabolic network:
Reaction Fluxes
The i-th reaction has a flux (rate) vi that is flowing.
vi multiplies each metabolite in the i-th reaction.
48. Matrix representation of a metabolic network
11 metabolites and 4 reactions
Use S ∈ Rm×n and flux vector v ∈ Rn to express the
change of the mass of each metabolite over time [Palsson’15],
dr
dt
= Sv = S
v1
v2
v3
v4
= v1S(·,1) + v2S(·,2) + v3S(·,3) + v4S(·,4)
49. Steady states: The network in balance
When for each metabolite, the rate of production equals to the
rate of consumption, the reactions exactly balance each other.
When a flux vector v balances the network,
Sv = 0,
v is a steady state.
50. The region of steady states
As a low dimensional polytope.
Sv = 0,
vlb ≤ v ≤ vub ←→
v=Nx
S ∈ Rm×n
, v ∈ Rn
As a full dimensional polytope
P := {x ∈ Rd | Ax ≤ b}
N ∈ Rn×d the matrix of the right nullspace of S.
51. Sampling steady states
Sampling could lead to important biological insights [Palsson’15].
Explore the flux space [Schellenberger,Palsson’09].
We introduce a
Multiphase Monte Carlo Sampling algorithm
based on Billiard Walk
52. Difficulties - skinny polytopes
When the polytope is skinny:
The average number of reflections increases.
The mixing rate decreases.
53. Multiphase Monte Carlo Sampling
MMCS(P0, p, N, set i = 0)
1 Sample O(d) points from Pi with Billiard walk.
2 Estimate the Effective Sample Size (ESS) of the sample
in Pi .
3 Map the sampled points to an isotropic position and
apply the same transformation Ti to Pi , set
Pi+1 = Ti (Pi ).
4 i = i + 1; goto 1.
5 Stop when the sum of ESS N and PSRF 1.1.
Related work
(optimization)
1. [Bertsimas et al.’04]
2. [Kalai et al.’06]
55. Find possible anti-COVID19 targets
Not an anti-viral target Possible anti-viral target
Sample steady states when,
the growth rate of COVID-19 is optimized,
the host biomass production is optimized.
Check if the flux distribution of a reaction changes
[Renz,Widerspick,Dräger’20,’21].
[Open-source internship
Tweag’21]
Joint work:
[Tweag] Cheplyaka, Carstens
[GeomScale] Fisikopoulos, Tsigaridas,
ii Zafeiropoulos
57. Complexity
Computing the exact volume of P,
is #P-hard for all the representations [DyerFrieze’88]
is open if both H- and V- representations available
is APX-hard (oracle model) [Elekes’86]
58. Randomized approximation algorithms
Multiphase Monte Carlo
Theorem [Dyer, Frieze, Kannan’91]
For any convex body P and any 0 ≤ , δ ≤ 1, there is a
randomized algorithm which computes an estimate V s.t. with
probability 1 − δ we have (1 − )vol(P) ≤ V ≤ (1 + )vol(P),
and the number of oracle calls is poly(d, 1/, log(1/δ)).
Using randomness, we can go from an exponential
approximation to an arbitrarily small one.
59. State-of-the-art
Authors-Year
Complexity
random walk
(oracle calls)
[Dyer, Frieze, Kannan’91] O∗(d23) grid walk
[Kannan, Lovasz, Simonovits’97] O∗(d5) ball walk
[Lovasz, Vempala’03] O∗(d4) hit-and-run
[Cousins, Vempala’15] O∗(d3) ball walk
Can not be implemented as they are due to large constants in
the complexity and pessimistic theoretical bounds.
Practical algorithms:
Follow theory but make practical adjustments (experimental).
[Emiris, Fisikopoulos’14] Sequence of balls + coordinate
hit-and-run.
[Cousins, Vempala’16] Spherical Gaussians + hit-and-run
60. Practical Multiphase Monte Carlo scheme
Let Cm ⊆ · · · ⊆ C1 a sequence of scaled copies of a body C
intersecting P, s.t. Cm ⊆ P ⊆ C1 [C,Emiris,Fisikopoulos’21].
vol(P) = vol(P ∩ Cm) · vol(P∩Cm−1)
vol(P∩Cm)
· · · vol(P∩C1)
vol(P∩C2)
· vol(P)
vol(P∩C1)
61. Ratio estimation
Estimate ri = vol(P∩Ci+1)
vol(P∩Ci )
within some target relative error i .
Sample N uniform points from Pi = Ci ∩ P and count points
in Pi+1 = Ci+1 ∩ P ⊆ Pi .
Keep each ratio bounded, then N = O(1/2
i ) points suffices.
Use Billiard walk to sample uniformly from each body.
62. Keep volume ratios bounded
Statistical tests
Given convex bodies Pi ⊇ Pi+1, we define two statistical tests:
[U-test(Pi , Pi+1)] H0: vol(Pi+1)/vol(Pi ) ≥ r + δ
[L-test (Pi , Pi+1)] H0: vol(Pi+1)/vol(Pi ) ≤ r
The U-test and L-test are successful iff both H0 are rejected.
If both U-test and L-test are successful then
ri = vol(Pi+1)/vol(Pi ) ∈ [r, r + δ], with high probability.
63. Experiments
Total number of generated points: ∼ d2 (experimental).
Our method outperforms existing implementations
64. Experiments
We compute the volume of Birkhoff polytope for n 15 for
the first time.
Bn d Exact Vol. Asympt. Vol. Est. Vol m steps time (sec)
B10 81 8.78e-46 9.81e-46 8.56e-46 13.1 4.69e+04 1.9
B20 361 ?? 4.23e-312 3.31e-312 94.9 8.46e+05 297
B25 576 ?? 6.46e-554 5.56e-554 158.8 1.82e+06 1 060
B30 841 ?? 1.60e-875 1.35e-875 249.1 3.53e+06 4 982
B33 1024 ?? 1.56e-1108 1.37e-1108 314.1 4.97e+06 11 105
We perform computations up to thousands of dimensions for
the first time.
Polytope Vol. error m steps time
cube-500 3.38e+150 0.03 84.7 7.09e+05 458
cube-1000 2.11e+301 0.03 201.0 2.70e+06 6 180
simplex-500 7.55e-1135 0.08 149.1 1.65e+06 847.8
simplex-1000 2.67e-2568 0.07 328.4 5.34e+06 10 711
product-simplices-250-250 9.10e-986 0.05 151.3 1.69e+06 958.5
product-simplices-500-500 6.22e-2269 0.07 332.5 5.39e+06 9 432
iSDY 1059 [2966-509] 2.25e-350 ?? 136.0 1.43e+06 3 486
Recon1 [4934-931] 8.01e-5321 ?? 266.1 3.84e+06 18 225
66. Crisis detection
Financial markets exhibit 3 types of behavior e.g.,
[Billio, Getmansky,Pelizzon’12]:
1 In normal times, stocks are characterized by slightly positive
returns and a moderate volatility.
2 In up-market times (typically bubbles) by high returns and low
volatility.
3 During financial crises by strongly negative returns and high
volatility.
These observations motivate us to describe the time-varying
dependency between portfolios’ returns and volatility.
67. Geometric representation of portfolios
Set of portfolios
{x ∈ Rd
|
P
i xi = 1, xi ≥ 0}
The set of portfolios can be seen
as the simplex/polytope in Rd
.
Ptf x= (0.3, 0.25, 0.45)
69. Computing the copula of a given time period
– Cut the simplex with bothqqqa
q hyperplanes and ellipsoids
– Compute the volume of the q
d bodies defined by the qdfxcvxc
d intersections
→
– Obtain a Copula
– Bivariate distribution
– Each marginal is uniform
70. Q: How to detect a crisis?
1st September 1999 1st September 2000
during dot-com bubble bubble burst
Left Copula corresponds to normal and right copula to crises times.
⇒ Normal times: The mass of portfolios on the up diagonal.
⇒ Crisis times: The mass of portfolios on the down diagonal.
71. Answer: Define an indicator I
I :=
mass in red area
mass in blue area
72. Detects past financial crises
Daily returns from Europe DJ 600 from 01/01/1990 to 31/11/2017
I 1 for 61-100 days , over 100 days
[Calès,C,Emiris,Fisikopoulos’18]
1. May ’90 - Dec. ’90: early 90’s recession.
2. May ’00 - May ’01: dot-com bubble burst.
3. Oct. ’01 - Apr. ’02: the stock market downturn of 2002.
4. Nov. ’05 - Apr. ’06: not listed.
5. Dec. ’07 - Aug. ’08: recent crisis.
73. Portfolio strategies
Portfolio managers compute and propose portfolio allocations.
An investor decides which asset allocation proposal to select
and how much to modify it.
To model this procedure, we employ log-concave distributions
74. Portfolio strategies
Let π be an LC distribution supported on the portfolio domain,
Then, Fπ is a portfolio allocation strategy:
“To build a portfolio with strategy Fπ sample a
point/portfolio from π”
75. Mixed strategy
Let π1, . . . , πM be a sequence of LC distributions.
We call Fπ the mixed strategy;
π(x) =
PM
i=1 wi πi (x) a mixture density;
wi ≥ 0,
PM
i=1 wi = 1.
Weight wi the proportion of the investors that build their
portfolios according to Fπi .
76. A new portfolio score
For given asset returns R ∈ Rd over a single period of time, the
score of a portfolio, providing a value of return R∗, is
s =
Z
P
g(x)π(x)dx, g(x) =
1. if RT x ≤ R∗,
0, otherwise.
(3)
The score s corresponds to the expected proportion of
portfolios that an allocation outperforms when the portfolios
are invested according to the mixed strategy Fπ.
Theorem. We can estimate the score within accuracy after
O∗(Md4/2) operations.
77. Portfolio strategies - special case
equal risk different risk
πα,q ∝ e−αφq(x), φ = xT Σx − qµT x ∗
µ: mean of assets’ returns
Σ: covariance matrix of assets’ returns
α: controls dispersion around manager’s proposal
q: controls the risk of the investment
∗Quadratic utility function in original Markowitz’s problem [’79,’84]; Nobel Prize in economics ’90.
78. Parametric score
[C,Christoforou,Dalamagas,Emiris’21]
We introduce behavioral functions to model how the investors
tend to behave (set the parameters α and q and the bias
vector r).
Parametric score, s(T) =
R
S hπ(w)e(rT w)/T , T 0.
T → ∞: investors are equally divided among strategies.
T → 0: all investors follow a single strategy.