A numerical analysis approach to convex optimization Speaker
- Slides: 128
A numerical analysis approach to convex optimization Speaker Rasmus Kyng ETH Zurich Joint works with D. Adil, R. Peng, S. Sachdeva, D. Wang March, 2020
The problem
The problem
The problem
The problem
The problem
The problem
The problem
Convex optimization, high accuracy slow running time fast running time Inf-norm regression 2 -norm regression Karmarkar; Chinese mathematicians, 2 nd century Renegar; Ye; Nesterov & Nemirovski Gauss, 19 th century
Convex optimization, high accuracy slow running time fast running time Nesterov & Nemirovski ‘ 94 p-norm regression Inf-norm regression 2 -norm regression
Convex optimization, high accuracy slow running time fast running time Bubeck, Cohen, Lee, Li ‘ 18 p-norm regression Nesterov & Nemirovski ‘ 94 p-norm regression Inf-norm regression 2 -norm regression
Our result compared to BCLL Our result
Our result compared to BCLL Our result
Convex optimization, high accuracy slow running time fast running time Adil, K. , Peng, Sachdeva ‘ 19 p-norm regression Bubeck, Cohen, Lee, Li ‘ 18 p-norm regression Inf-norm regression 2 -norm regression
Convex optimization, high accuracy slow running time fast running time Adil, K. , Peng, Sachdeva ‘ 19 p-norm regression Bubeck, Cohen, Lee, Li ‘ 18 p-norm regression Inf-norm regression 2 -norm regression “Smooth” convex problems? !
Convex optimization, high accuracy slow running time fast running time Lee & Sidford ‘ 15 Inf-norm regression, w/ inverse maintenance Inf-norm regression Inverse maintenance: Karmarkar; Nesterov & Nemirovski; Vaidya; Lee & Sidford 2 -norm regression p-norm regression w/ inverse maintenance Adil, K. , Peng, Sachdeva ‘ 19 Inf-norm regression, w/ inverse maintenance Cohen, Lee, Song ‘ 19
Convex optimization, high accuracy slow running time fast running time 2 -norm regression p-norm regression w/ inverse maintenance Inf-norm regression, w/ inverse maintenance
Convex optimization, high accuracy slow running time fast running time Bubeck, Cohen, Lee, Li ’ 18 p-norm flow 2 -norm regression p-norm regression w/ inverse maintenance Inf-norm regression, w/ inverse maintenance inf-norm flow aka maxflow Goldberg-Rao `98 2 -norm flow Spielman Teng ‘ 04
Convex optimization, high accuracy slow running time fast running time Bubeck, Cohen, Lee, Li ’ 18 p-norm flow 2 -norm regression inf-norm flow aka maxflow 2 -norm flow Adil, K. , Peng, Sachdeva ‘ 19 p-norm regression w/ inverse maintenance Inf-norm regression, w/ inverse maintenance p-norm flow most flows? !
Convex optimization, high accuracy slow running time fast running time unweighted maxflow Madry ‘ 13 2 -norm regression p-norm regression w/ inverse maintenance Inf-norm regression, w/ inverse maintenance inf-norm flow aka maxflow 2 -norm flow unweighted (large) p-norm flow K. , Peng, Sachdeva, Wang ‘ 19 most flows? !
Convex optimization, high accuracy slow running time fast running time unweighted maxflow Liu-Sidford ‘ 2020 2 -norm regression p-norm regression w/ inverse maintenance Inf-norm regression, w/ inverse maintenance inf-norm flow aka maxflow 2 -norm flow unweighted (large) p-norm flow K. , Peng, Sachdeva, Wang ‘ 19 most flows? !
Convex optimization
2 nd order methods local upper bound (quadratic) local lower bound (quadratic)
Homotopy: making 2 nd order methods global Solve a sequence of different optimization problems, ensure starting point near next optimum
Homotopy: making 2 nd order methods global Solve a sequence of different optimization problems, ensure starting point near next optimum
Homotopy: making 2 nd order methods global Solve a sequence of different optimization problems, ensure starting point near next optimum
Homotopy: making 2 nd order methods global Solve a sequence of different optimization problems, ensure starting point near next optimum
local upper bound (quadratic) local lower bound (quadratic)
global upper bound (? ? ) global lower bound (? ? )
global upper bound (? ? ) global lower bound (? ? ) What makes our “envelopes” tractable?
global upper bound (? ? ) global lower bound (? ? ) What makes our “envelopes” tractable?
global upper bound (? ? ) global lower bound (? ? )
The ingredients of an algorithm
The envelopes (upper and lower bounds)
Good envelopes?
Good envelopes? Step progress
Good envelopes? Step progress Mult. weight oracle
Good envelopes!
Good envelopes! linear apx.
Good envelopes! linear apx. local curvature
Good envelopes! linear apx. local curvature "long-range behavior” Enough to give a global approximation!
linear apx. "long-range behavior” local curvature
lower bound upper bound Step progress Mult. weight oracle
Good envelopes! Long range: 4 th power
Good envelopes! Long range: 4 th power Locally quadratic
Iterative refinement with multiplicative weight method step problem Multiplicative weight method
Using multiplicative weights to take a step
Crude solver via multiplicative weights
Multiplicative weights refresher
But when is the oracle bad?
But when is the oracle bad? 1 unit on every edge
But when is the oracle bad?
But when is the oracle bad?
The dirty secrets
Crude solver via multiplicative weights
Why do we call it “iterative refinement”?
Iterative refinement step problem A. k. a. iterative refinement
Iterative refinement for linear equations step problem A. k. a. iterative refinement
Iterative refinement For linear equations, build the “crude” solver by preconditioning For our non-linear equations build the “crude” solver by multiplicative weights Or we can build a solver by recursive preconditioning Must be nonlinear, adaptive?
Recap [Adil-K. -Peng-Sachdeva]
Going much further for flows
Going much further for flows
Iterative refinement step problem A. k. a. iterative refinement
Iterative refinement with multiplicative weight method step problem Multiplicative weight method
Recursive Preconditioning? step problem Recurse! Solved by recursion
Recursive Preconditioning? step problem Recurse! Solved by recursion
The recursion works out
Smaller instances: ultrasparsification
Expanders
Expanders
Expanders
Expanders What are the flow maps?
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification
Repeat Smaller instances: Ultrasparsification Preserve linear, quadratic, and pth power simultaneously? Ensure projection of gradient to partition cycle space is uniform.
Repeat Smaller instances: Ultrasparsification Non-linear gaussian elimination? It helps that we are dealing with a circulation!
Recap
Application to maximum flow
But when is the oracle bad?
But when is the oracle bad?
But when is the oracle bad? Reweigh
But when is the oracle bad?
Open questions Remove unit weight restriction Generalize to most convex flow problems? Maximum flow? (iterative refinement works) Preconditioning outside graph-land?
Thanks! My co-authors Deeksha Adil Richard Peng Sushant Sachdeva UToronto Ga. Tech UToronto Di Wang Ga. Tech
Folklore (Spielman, Madry)
1 unit on every edge
Thanks! My co-authors Deeksha Adil Richard Peng Sushant Sachdeva UToronto Ga. Tech UToronto Di Wang Ga. Tech
TODO Reconcile m and choices? Decision on bold vs non-bold vectors Introduce flow problems somewhere Be clearer about additive vs multiplicative error! Be clear about recursion errors
- Pinwheels and polygons
- Convex hull is the smallest convex set
- Convex optimization in machine learning javatpoint
- Exact matrix completion via convex optimization
- Convex optimization
- Numerical optimization techniques for engineering design
- Graphical and numerical methods
- Definition interpolation
- Numerical analysis formula
- Finite differences and interpolation
- Types of errors in numerical analysis
- Newton's forward interpolation formula
- Secant method solved examples
- C programming and numerical analysis an introduction
- Trapezoidal riemann sum
- Numerical analysis
- Linear optimization and prescriptive analysis
- Network configuration optimization analysis
- Datagram approach and virtual circuit approach
- Theoretical models of counseling
- Waterfall and shower strategy
- Approach approach conflict
- Bandura's reciprocal determinism
- Research approach example
- Approach to system development
- Deep learning approach and surface learning approach
- Convex contour symbol
- Convex limacon
- Convex lens ray diagram
- Polygons definition
- Which of these are concave polygon
- Poligonul regulat cu 4 laturi se numeste
- Dimpled limacon
- Convex vs concave teeth
- Mirror lens equation
- Concave vs convex light refraction
- Salt of concave mirrors
- Astimatigma
- Concave lens equation
- Grades of joint mobilization
- Convex curve
- Property of converging lens
- Limacon with dimple
- Decide whether the figure is a polygon
- Rom elbow
- Diffraction by circular aperature
- V principle of growth
- Graham scan
- Converging lens in water
- Convex lens are thickest at the middle
- Kanat tangwongsan
- Concave vs convex polygon
- Convex vs concave lens
- Convex lens table
- Convex hull collision detection
- Salt for concave mirrors
- Lobate bacteria
- Convex polygon examples
- Is flat, smooth mirror
- Focal length formula
- Not polygons
- Concave mirror matter
- What is the sum of a decagon
- Convex
- Image formed by convex mirror when object is at infinity
- Incremental convex hull algorithm
- Geometric optics in everyday life
- Convex mirror is a diverging mirror
- Formula concave mirror
- Thin lens equation height
- Ray tracing convex lens
- Constrictive population pyramid
- Convex population pyramid
- Formule poligoane regulate
- Convex 39m 60m
- Converging lens shape
- Square edge convex
- Convex
- Regular reflection
- Concave and convex mirror
- Medium sized nose
- Types of polygons
- Polygon or not
- Uses of concave mirror
- Angle convex i concau
- Cermin cembung convex biasa disebut cermin negatif karena
- Shape of quadrilateral
- Ray diagram of concave
- Convex hull
- Is a rhombus a square
- Closed figure
- Peripheral joint
- 180 degree angle
- Concave lens cases
- Photographic enlarger ray diagram
- Mirror that curves outward and use in convenience store
- Curved
- Mehran convex mirror
- Sqdc
- Uses of concave and convex mirror
- Uses of convex mirror
- Image formation
- Sum of interior angles of a polygon
- 12 lead 심전도
- The rearview mirror should be checked
- Convex and concave polygon in computer graphics
- Numerical
- Numerical stroop
- The angle of acceptance cone is twice the
- Ringkasan numerik adalah
- Counterdiscipline
- Numerical integration excel
- Basic algebra definition
- What is numerical pattern
- Numerical computing with python
- Numerical differentiation c++
- 1st order derivative formula
- What is the lower quartile measure of this box plot?
- Numerical descriptive measures exercises
- Taylor series numerical methods
- Medical
- Dark field microscopy
- What are numerical expressions
- Numerical data examples
- Numerical aperture in microscope
- Visual numerical learning style
- Numerical flux
- Fleet numerical meteorology and oceanography center
- Turner syndrome is what numerical chromosome disorder?