九大数理集中講義 Comparison, Analysis, and Control of Biological Networks (5) Control of Probabilistic Boolean Networks Tatsuya Akutsu Bioinformatics Center Institute.

Slides:



Advertisements
Similar presentations
Bayesian Belief Propagation
Advertisements

CS188: Computational Models of Human Behavior
Minimum Clique Partition Problem with Constrained Weight for Interval Graphs Jianping Li Department of Mathematics Yunnan University Jointed by M.X. Chen.
A Hierarchical Multiple Target Tracking Algorithm for Sensor Networks Songhwai Oh and Shankar Sastry EECS, Berkeley Nest Retreat, Jan
Bayesian Networks CSE 473. © Daniel S. Weld 2 Last Time Basic notions Atomic events Probabilities Joint distribution Inference by enumeration Independence.
A Novel Method For Fast Model Checking Project Report.
Dynamic Bayesian Networks (DBNs)
Interchanging distance and capacity in probabilistic mappings Uriel Feige Weizmann Institute.
02/01/11CMPUT 671 Lecture 11 CMPUT 671 Hard Problems Winter 2002 Joseph Culberson Home Page.
Complexity Theory CSE 331 Section 2 James Daly. Reminders Project 4 is out Due Friday Dynamic programming project Homework 6 is out Due next week (on.
Regulatory Network (Part II) 11/05/07. Methods Linear –PCA (Raychaudhuri et al. 2000) –NIR (Gardner et al. 2003) Nonlinear –Bayesian network (Friedman.
1 Optimization problems such as MAXSAT, MIN NODE COVER, MAX INDEPENDENT SET, MAX CLIQUE, MIN SET COVER, TSP, KNAPSACK, BINPACKING do not have a polynomial.
The Theory of NP-Completeness
CSE 421 Algorithms Richard Anderson Lecture 27 NP Completeness.
Chapter 11: Limitations of Algorithmic Power
6. Gene Regulatory Networks
Profile Hidden Markov Models PHMM 1 Mark Stamp. Hidden Markov Models  Here, we assume you know about HMMs o If not, see “A revealing introduction to.
1 Integrality constraints Integrality constraints are often crucial when modeling optimizayion problems as linear programs. We have seen that if our linear.
Dependency networks Sushmita Roy BMI/CS 576 Nov 26 th, 2013.
Introduction to Profile Hidden Markov Models
九大数理集中講義 Comparison, Analysis, and Control of Biological Networks (3) Domain-Based Mathematical Models for Protein Evolution Tatsuya Akutsu Bioinformatics.
Approximation Algorithms for Stochastic Combinatorial Optimization Part I: Multistage problems Anupam Gupta Carnegie Mellon University.
Yossi Azar Tel Aviv University Joint work with Ilan Cohen Serving in the Dark 1.
九大数理集中講義 Comparison, Analysis, and Control of Biological Networks (7) Partial k-Trees, Color Coding, and Comparison of Graphs Tatsuya Akutsu Bioinformatics.
1 MCMC Style Sampling / Counting for SAT Can we extend SAT/CSP techniques to solve harder counting/sampling problems? Such an extension would lead us to.
Introduction to Job Shop Scheduling Problem Qianjun Xu Oct. 30, 2001.
Introduction to the Hankel -based model order reduction for linear systems D.Vasilyev Massachusetts Institute of Technology, 2004.
ECES 741: Stochastic Decision & Control Processes – Chapter 1: The DP Algorithm 31 Alternative System Description If all w k are given initially as Then,
Statistical Sampling-Based Parametric Analysis of Power Grids Dr. Peng Li Presented by Xueqian Zhao EE5970 Seminar.
Efficient Solution Algorithms for Factored MDPs by Carlos Guestrin, Daphne Koller, Ronald Parr, Shobha Venkataraman Presented by Arkady Epshteyn.
Attractor Detection and Control of Boolean Networks Tatsuya Akutsu Bioinformatics Center Institute for Chemical Research Kyoto University.
九大数理集中講義 Comparison, Analysis, and Control of Biological Networks (4) Analysis and Control of Boolean Networks Tatsuya Akutsu Bioinformatics Center Institute.
CSC 172 P, NP, Etc. “Computer Science is a science of abstraction – creating the right model for thinking about a problem and devising the appropriate.
Introduction to Bayesian Networks
NP-COMPLETENESS PRESENTED BY TUSHAR KUMAR J. RITESH BAGGA.
EMIS 8373: Integer Programming NP-Complete Problems updated 21 April 2009.
Resource Mapping and Scheduling for Heterogeneous Network Processor Systems Liang Yang, Tushar Gohad, Pavel Ghosh, Devesh Sinha, Arunabha Sen and Andrea.
1 The Theory of NP-Completeness 2 Cook ’ s Theorem (1971) Prof. Cook Toronto U. Receiving Turing Award (1982) Discussing difficult problems: worst case.
CS774. Markov Random Field : Theory and Application Lecture 02
Markov Chains and Random Walks. Def: A stochastic process X={X(t),t ∈ T} is a collection of random variables. If T is a countable set, say T={0,1,2, …
AAAI 2011, San Francisco Trajectory Regression on Road Networks Tsuyoshi Idé (IBM Research – Tokyo) Masashi Sugiyama (Tokyo Institute of Technology)
Gibbs sampling for motif finding Yves Moreau. 2 Overview Markov Chain Monte Carlo Gibbs sampling Motif finding in cis-regulatory DNA Biclustering microarray.
1/19 Minimizing weighted completion time with precedence constraints Nikhil Bansal (IBM) Subhash Khot (NYU)
CSE 421 Algorithms Richard Anderson Lecture 27 NP-Completeness and course wrap up.
Topics in Algorithms 2007 Ramesh Hariharan. Tree Embeddings.
Bioinformatics Center Institute for Chemical Research Kyoto University
1 CS612 Algorithms for Electronic Design Automation CS 612 – Lecture 8 Lecture 8 Network Flow Based Modeling Mustafa Ozdal Computer Engineering Department,
1 An Arc-Path Model for OSPF Weight Setting Problem Dr.Jeffery Kennington Anusha Madhavan.
CPS Computational problems, algorithms, runtime, hardness (a ridiculously brief introduction to theoretical computer science) Vincent Conitzer.
Stochastic Optimization for Markov Modulated Networks with Application to Delay Constrained Wireless Scheduling Michael J. Neely University of Southern.
1 CPSC 320: Intermediate Algorithm Design and Analysis July 30, 2014.
1 Relational Factor Graphs Lin Liao Joint work with Dieter Fox.
Space Complexity. Reminder: P, NP classes P is the class of problems that can be solved with algorithms that runs in polynomial time NP is the class of.
1 Structure Learning (The Good), The Bad, The Ugly Inference Graphical Models – Carlos Guestrin Carnegie Mellon University October 13 th, 2008 Readings:
Instructor: Shengyu Zhang 1. Optimization Very often we need to solve an optimization problem.  Maximize the utility/payoff/gain/…  Minimize the cost/penalty/loss/…
Probabilistic Robotics Probability Theory Basics Error Propagation Slides from Autonomous Robots (Siegwart and Nourbaksh), Chapter 5 Probabilistic Robotics.
Krishnendu ChatterjeeFormal Methods Class1 MARKOV CHAINS.
Motoki Shiga, Ichigaku Takigawa, Hiroshi Mamitsuka
Profile Hidden Markov Models PHMM 1 Mark Stamp. Hidden Markov Models  Here, we assume you know about HMMs o If not, see “A revealing introduction to.
CS4234 Optimiz(s)ation Algorithms L2 – Linear Programming.
Lecture 7: Constrained Conditional Models
Independent Cascade Model and Linear Threshold Model
Maximum Expected Utility
Non-additive Security Games
Exact Inference Continued
Intro to Theory of Computation
Hidden Markov Models Part 2: Algorithms
Chapter 11 Limitations of Algorithm Power
Bioinformatics Center Institute for Chemical Research Kyoto University
Independent Cascade Model and Linear Threshold Model
Presentation transcript:

九大数理集中講義 Comparison, Analysis, and Control of Biological Networks (5) Control of Probabilistic Boolean Networks Tatsuya Akutsu Bioinformatics Center Institute for Chemical Research Kyoto University

Contents Boolean Network Probabilistic Boolean Network (PBN)  Formulation, Control of PBN Finding an Optimal Path in PBN  ILP-based method Minimizing the Maximum Cost in PBN  Optimize the worst case control cost Hardness of Control of PBN  Control of PBN is ∑ 2 p -hard Summary

Probabilistic Boolean Network

Probabilistic Boolean Network (PBN) Multiple control rules (boolean functions) for each node Control rule is selected randomly at each t according to a given probability distribution  Almost equivalent to Dynamic Bayesian Network  Pros: Capable of noise. Can be modeled as Markov process.  Cons : Not scalable since it takes O(2 n ) or more time for almost all problems on PBN A B C A(t+1) = B(t) AND C(t) A(t+1) = B(t) OR (NOT C(t)) with Prob.=0.6 with Prob.=0.4 [Shmulevich et al., 2002]

Example of PBN PBN State Transition Diagram (only for half of nodes) One of 4(=2×1×2) BNs is randomly selected at each time step

BN vs. PBN BN: 1 outgoing edge PBN: multiple outgoing edges (with probabilities) BNPBN BN 1 BN 2 BN 3 BN

BN vs. PBN in State Transition Diagram BN: 1 outgoing edge ⇔ PBN: multiple outgoing edges (with probabilities) BN PBN

Matrix Representation of a BN A ’ B ’ C ’ A B C State transition can be written as where State transition table

Matrix Representation of a PBN The above means:

PBN-CONTROL: Model Probabilistic Boolean network (PBN, an extension of Boolean network) Global state at time t : Probabilistic regulation rule is given as a 2 n ×2 n matrix A A can be controlled by m boolean variables Cost functions  C t (v, u) : cost for applying control u for global state v at time t  C(v) : cost for final global state v [Datta et al., Machine Learning, 2003]

PBN-CONTROL: Problem and Algorithm Problem:  Given initial state v(0), control rule A(u(t)), target time M, and cost functions,  Find a first control action u(0) minimizing Can be solved by dynamic programming [Datta et al., Machine Learning, 2003]

PBN-CONTROL: Dynamic Programming But, it takes too long CPU time because A is 2 n ×2 n matrix

Finding an Optimal Path in PBN

Problem: Finding a control sequence for a PBN with the maximum probability Method: Introduction of variable y r,t y r,t =1 iff r -th BN is selected at time step t Modification of ILP p r : probability of selecting r -th BN We also developed DP (Dynamic Programming) algorithm (with exponential time and space) [Chen et al., IEEE BIBM, 2010]

Optimal Path Problem t=0 t=1 t=2 Control v(0) v 1 (1)v 2 (1)v 3 (1) u2u2 v 1 (2) u2u2 u1u Optimal Path Control (u2,u1)(u2,u1) u1u u1u1 0.4 Pr(BN 1 )=0.4 Pr(BN 2 )=0.6 v i, u j : global state

Minimizing the Maximum Cost in PBN

Minimizing the Maximum Cost (1) Problem: Given initial state v(0), control rule A(u(t)), target time M, and cost functions, find a first control action u(0) minimizing Minimizing the worst case cost is important Method: dynamic programming (with exponential time/space) F(v t,u t ) : set of states that can be reached from v t with control u t [Chen et al., IEEE BIBM, 2010]

Minimizing the Maximum Cost (2) Pr(BN 1 )=0.4 Pr(BN 2 )=0.6 v(0) v 1 (1)v 2 (1)v 3 (1) u2u2 0.4 u1u1 u1u1 u2u2 u1u1 u2u2 0.6 u1u1 u2u2 v 1 (2) CM(v)CM(v) like minimax game

Minimizing the Average Cost v(0) v 1 (1)v 2 (1)v 3 (1) u2u2 0.4 u1u1 u1u1 u2u2 u1u1 u2u2 0.6 u1u1 u2u2 v 1 (2) Original Control Problem for PBN weighted sum of costs max +

Hardness Results

Limitation of Use of ILP ILP can be used for  Control of BN  Finding an optimal path in PBN But, cannot be used for  Control of PBN Minimizing the average cost  Minimizing the maximum cost Why ?

Hardness Results (1) Thm. 1: Minimizing the maximum cost in control of PBN is -hard. Thm. 2: Minimizing the average cost in control of PBN (i.e., original control problem for PBN) is -hard. Proof: Reduction from QBF for 3-DNF Implications Control of PBN is harder than Control of BN (NP-complete [Akutsu et al, 2007] ) because is believed to be harder than NP Such technique as ILP, SAT cannot be utilized because these are in NP [Chen et al., IEEE BIBM, 2010]

Hardness Results (2) Control of BN is NP-complete (for fixed time steps) Integer linear programming (ILP)-based method for control of BN Control of PBN is harder than NP ( -hard) Such technique as ILP, SAT cannot be utilized [Akutsu et al., JTB 07] [Akutsu et al., IEEE CDC 09]

Summary

Probabilistic Boolean Network  Probabilistic extension of BN  DP algorithm for control of PBN Finding an Optimal Path in PBN  ILP-based method Minimizing the Maximum Cost  DP algorithm Hardness of Control of PBN  Harder than NP and Control of BN