230 likes | 314 Views
Factor Graphs. 2005. 5. 20 Young Ki Baik Computer Vision Lab. Seoul National University. Contents. Introduction Sum product algorithm Computing a single marginal function Computing all marginal function Probabilistic modeling Conclusion. Introduction. What is a Factor Graph?
E N D
Factor Graphs 2005. 5. 20 Young Ki Baik Computer Vision Lab. Seoul National University
Contents • Introduction • Sum product algorithm • Computing a single marginal function • Computing all marginal function • Probabilistic modeling • Conclusion
Introduction • What is a Factor Graph? • A factor graph shows how a function of several variables can be factored into a product of “smaller” function.
Introduction • What is a Factor Graph?
Introduction • Why are factor graphs useful? • Factor graphs simplify problems. • Many efficient algorithms can be applied to factor graphs. • Special Feature • Factor graph represent not only variables or constant, but also functions.
Sum product algorithm • Marginal function • “~{x}” notation • Marginal function Object : Get marginal function g(x) using factor graph
Sum product algorithm • Example (Simple Factor Graph) • Let be a function of four variables.
Computing a single marginal function • Example (Simple Factor Graph) • Marginal function for
Computing a single marginal function • Example (Simple Factor Graph) • Marginal function for
Computing a single marginal function • Example (Simple Factor Graph) • Marginal function for and Bottom-up procedure
Computing all marginal functions • Computing all marginal functions problem • In order to compute all marginal functions , we need to calculate single marginal function as much as n times. • Message passing algorithm • Solution for redundancy problem
Computing all marginal functions • Message passing algorithm • Let denote the message sent from node x to node f in the operation. • Let denote the message sent from node f to node x.
Computing all marginal functions • Message passing algorithm • Variable to local function • Local function to variable
Computing all marginal functions • A Detail Example (Message passing algorithm) • The message may be generated in 4 steps. • ➀ ~ ➃ are step of message passing algorithm • Algorithms start from each leaves. ➃ ➂ ➀ ➁ ➀ ➁ ➂ ➃ ➂ ➃ ➁ ➀
Computing all marginal functions • A Detail Example (Message passing algorithm) • Step 1: Variable to local function ➀ ➀ ➀
Computing all marginal functions • A Detail Example (Message passing algorithm) • Step 2: Local function to variable ➁ ➁ ➁
Computing all marginal functions • A Detail Example (Message passing algorithm) • Step 3: Variable to local function ➂ ➂ ➂
Computing all marginal functions • A Detail Example (Message passing algorithm) • Step 4: Local function to variable ➃ ➃ ➃
Computing all marginal functions • A Detail Example (Message passing algorithm) • Termination ➃ ➂ ➀ ➁ ➀ ➁ ➂ ➃ ➂ ➃ ➁ ➀
Probabilistic Modeling • Markov chain
Probabilistic Modeling • Hidden markov model
Conclusion • More information • The closed factor graph • It can be computed by iterative method. • Conclusion • Factor graph can apply to many efficient algorithms. • Factor graph is only a simplifying tool to solve the problems.