Non-commutative computation with division Avi Wigderson IAS, Princeton Pavel Hrubes U. Washington.

Slides:



Advertisements
Similar presentations
Rules of Matrix Arithmetic
Advertisements

5.4. Additional properties Cofactor, Adjoint matrix, Invertible matrix, Cramers rule. (Cayley, Sylvester….)
3.2 Determinants; Mtx Inverses
CHAPTER ONE Matrices and System Equations
CS151 Complexity Theory Lecture 5 April 13, 2004.
Arithmetic Hardness vs. Randomness Valentine Kabanets SFU.
ECIV 301 Programming & Graphics Numerical Methods for Engineers Lecture 14 Elimination Methods.
ECIV 520 Structural Analysis II Review of Matrix Algebra.
ECIV 301 Programming & Graphics Numerical Methods for Engineers REVIEW II.
Copyright © Cengage Learning. All rights reserved. 7.6 The Inverse of a Square Matrix.
Linear Algebra - Chapter 1 [YR2005]
Linear Algebra With Applications by Otto Bretscher. Page The Determinant of any diagonal nxn matrix is the product of its diagonal entries. True.
Compiled By Raj G. Tiwari
Math 3121 Abstract Algebra I Lecture 3 Sections 2-4: Binary Operations, Definition of Group.
1 February 24 Matrices 3.2 Matrices; Row reduction Standard form of a set of linear equations: Chapter 3 Linear Algebra Matrix of coefficients: Augmented.
Polynomials Algebra Polynomial ideals
 Row and Reduced Row Echelon  Elementary Matrices.
Rev.S08 MAC 1140 Module 10 System of Equations and Inequalities II.
Systems of Linear Equation and Matrices
Chapter 7 Notes Honors Pre-Calculus. 7.1/7.2 Solving Systems Methods to solve: EXAMPLES: Possible intersections: 1 point, 2 points, none Elimination,
Multilinear NC 1  Multilinear NC 2 Ran Raz Weizmann Institute.
Recent Developments in Algebraic Proof Complexity Recent Developments in Algebraic Proof Complexity Iddo Tzameret Tsinghua Univ. Based on Pavel Hrubeš.
MAT 2401 Linear Algebra 2.3 The Inverse of a Matrix
7.4. Computations of Invariant factors. Let A be nxn matrix with entries in F. Goal: Find a method to compute the invariant factors p 1,…,p r. Suppose.
Overview Definitions Basic matrix operations (+, -, x) Determinants and inverses.
Matrices. A matrix, A, is a rectangular collection of numbers. A matrix with “m” rows and “n” columns is said to have order m x n. Each entry, or element,
Copyright © 2011 Pearson Education, Inc. Rational Expressions Section P.6 Prerequisites.
Notes 2.4 –Real Zeros of Polynomial Functions
Mathematical foundationsModern Seismology – Data processing and inversion 1 Some basic maths for seismic data processing and inverse problems (Refreshement.
MAT 2401 Linear Algebra 2.1 Operations with Matrices
Integrating high-level constructs into programming languages Language extensions to make programming more productive Underspecified programs –give assertions,
Umans Complexity Theory Lectures Lecture 1a: Problems and Languages.
Section 1.4 Inverses; Rules of Matrix Arithmetic.
For real numbers a and b,we always have ab = ba, which is called the commutative law for multiplication. For matrices, however, AB and BA need not be equal.
Linear Algebra Chapter 2 Matrices.
Happy 80th B’day Dick.
Ankit Garg Princeton Univ. Joint work with Leonid Gurvits Rafael Oliveira CCNY Princeton Univ. Avi Wigderson IAS Noncommutative rational identity testing.
Happy 60 th B’day Noga. Elementary problems encoding computational hardness Avi Wigderson IAS, Princeton or Some problems Noga never solved.
2.5 – Determinants and Multiplicative Inverses of Matrices.
Characterizing Propositional Proofs as Non-Commutative Formulas Iddo Tzameret Royal Holloway, University of London Based on Joint work Fu Li (Texas Austin)
Introduction to Financial Modeling MGT 4850 Spring 2008 University of Lethbridge.
Multiplicative Group The multiplicative group of Z n includes every a, 0
Linear Algebra Engineering Mathematics-I. Linear Systems in Two Unknowns Engineering Mathematics-I.
Boolean Algebra.
Recent Developments in Algebraic & Proof Complexity Recent Developments in Algebraic & Proof Complexity Iddo Tzameret Tsinghua Univ. Based on Hrubes and.
1 IAS, Princeton ASCR, Prague. The Problem How to solve it by hand ? Use the polynomial-ring axioms ! associativity, commutativity, distributivity, 0/1-elements.
Power Functions with Modeling power function: a function of the form Note: k and a can be any non-zero real number power: the number a constant of variation:
Linear Algebra With Applications by Otto Bretscher.
College Algebra Chapter 6 Matrices and Determinants and Applications
Algebraic Proofs over Noncommutative Formulas
Linear Algebra review (optional)
Linear Algebra Lecture 2.
7.7 Determinants. Cramer’s Rule
Part B. Linear Algebra, Vector Calculus
1.4 Inverses; Rules of Matrix Arithmetic
Chapter 0 Review of Algebra.
Matrix PI-algebras and Lower Bounds on Arithmetic Proofs (work in progress) Iddo Tzameret Joint work with Fu Li Tsinghua University.
Umans Complexity Theory Lectures
Proving Algebraic Identities
Proving Algebraic Identities
Hiroshi Hirai University of Tokyo
Elementary Matrix Methid For find Inverse
Linear Algebra in Weak Formal Theories of Arithmetic
Applied Discrete Mathematics Week 4: Functions
Matrices.
Elementary Linear Algebra
Linear Algebra review (optional)
MA5242 Wavelets Lecture 1 Numbers and Vector Spaces
Lecture 8 Matrix Inverse and LU Decomposition
CS151 Complexity Theory Lecture 5 April 16, 2019.
Presentation transcript:

Non-commutative computation with division Avi Wigderson IAS, Princeton Pavel Hrubes U. Washington

Arithmetic complexity – why? -Can’t deal with Boolean complexity -What can be computed with + − × ÷ ? -Linear algebra, polynomials, codes, FFT,… -Helps Boolean complexity (arithmetization) -………

Arithmetic complexity – basics X = (X) ij an n×n matrix. -Det n (X) = Σ σ sgn (σ) Π i X i σ (i)  “P” -Per n (X) = Σ σ Π i X i σ (i)  “NP” -(X) -1 : n 2 rational functions  “P” F field ÷ ÷ × × + + − − × × XiXi XjXj XiXi c + + S(f) – circuit size “P”: S is poly(n) L(f) – formula size “NC”: L is poly(n) n variables, f degree <n f

X 1, X 2,… commuting variables: X i X j = X j X i F[X 1, X 2,…] polynomial ring: p, q. F(X 1, X 2,… ) field of rational functions: pq -1 [Strassen’73] Division can be efficiently eliminated when computing polynomials (eg from Gauss elimination for computing Det). Since then, arithmetic complexity focused on , ,  We’ll restore division to its former (3 rd grade) glory! Commutative computation

State-of-the-art F[X 1,X 2,…] F  X 1, X 2,…  F(  X 1, X 2,…  ) comm, no ÷ non-comm, no ÷ non-comm Circuit lb Formula lb NC-hard NP-hard NC = P? P = NP? PIT (Word Problem) S> nlog n [BS] L> n 2 [K] Det [V] Per [V] P=NC [VSBR] Per n ≤ Det p(n) BPP [SZ,DL]

X 1, X 2,… non-commuting vars: X i X j  X j X i F  X 1, X 2,…  non-commut. polynomial ring: p, q. -Order of variables in monomials matter! E.g. Det n (X) = Σ σ sgn (σ) X 1 σ (1) X 2 σ (2)    X n σ (n) is just one option (Cayley determinant) -Weaker model. E.g. X 2 -Y 2 costs 2 multiplications, but just 1 in the commut. case: X 2 -Y 2 = (X-Y)(X+Y) Non-commutative computation (groups, matrices, quantum, language theory,…)

State-of-the-art F[X 1,X 2,…] F F{X 1,X 2,…} comm, no ÷ non-comm, no ÷ non-comm Circuit lb Formula lb NC-hard NP-hard NC = P? P = NP? PIT (Word Problem) S> nlog n [BS] L> n 2 [K] Det [V] Per [V] P=NC [VSBR] Per n ≤ Det p(n) ? BPP [SZ,DL] L(Det n )>2 n [N] Per [HWY] Det [AS] P  NC [N] BPP [AL,BW] L(X -1 )>2 n [HW] X -1 [HW] P  NC [HW] BPP?

The wonderful wierd world of non-commutative rational functions x −1 + y −1, yx −1 y have no expression fg −1 for polys f,g (x + xy −1 x) −1 = x −1 - (x + y) −1 Hua’s identity Can one decide equivalence of 2 expressions? (x + zy −1 w) −1 can’t eliminate this nested inversion! Reutenauer Thm: Inverting an nxn generic matrix requires n nested inversions. Key to the formula lower bound on X -1

The free skew field (I) [Amitsur] A “circuit complexity” definition! Field of fractions F(  X 1, X 2,…  ) of F  X 1, X 2,…  Take all formulae r(X 1, X 2,…) with , , , ÷ r~s if for all matrices M 1, M 2,…of all sizes r(M 1, M 2,…) = s(M 1, M 2,…) whenever they make sense (no zero division) Amitsur Thm: F(  X 1, X 2,…  ) is a skew field – every nonzero element is invertible! Word problem (RIT): Is r = 0?

The free skew field (II) [Cohn] Matrix inverse definition R an nxn matrix with entries in F  X 1, X 2,…  R is full if R ≠ AB with A n  r, B r  n, r<n. Ex: 0 X Y Singular if vars commute -X 0 Z Invertible if vars non-commut. -Y –Z 0 Cohn’s Thm: F(  X 1, X 2,…  ) is the field of entries of inverses of all full matrices over F  X 1, X 2,…  Key to formula completeness of X -1 Word problem: Is R invertible (full)? Cohn’s Thm: Decidable (via Grobner basis alg).

Minimal dimension problem Ex: 0 X Y Singular under M 1 (F)-substitutions -X 0 Z Invertible with M 2 (F) substitutions -Y –Z 0 Conjecture: Every full nxn R with entries in {X i }, F, is invertible under M d (F) substitutions, d=poly(n). - Conjecture true for polynomials [Amitsur-Levizky] - Conjecture implies: 1)RIT  BPP 2)Efficient elimination of division gates from non- commutative formulas computing polynomials 3)Degree bounds in Invariant Theory (& GCT )

÷