Madhu Sudan: Algebra and Computation

Course announcement

Preliminary course announcement

Number: 6.966
Title: Algebra and Complexity Theory
Instructor: Madhu Sudan
Meets: 2:30-4:00pm, MW in 34-302.

Prereq: Undergraduate courses in

  • Complexity theory (6.045) and
  • Discrete Math. (6.042) or Algebra (3.1415)

3-0-9 G-Level Grad Credit

This course studies the interplay between algebra and computation. The course will be divided into roughly three equal parts (1 <= #lectures/part <= 24).Algorithms for fundamental algebraic problems:

  • Factorization of polynomials:
    • Over reals, finite fields, integers;
    • For univariate and multivariate case;
    • Application to encoding+decoding error-correcting codes
  • Quantifier elimination: How to solve systems of polynomial equations. Application to kinematics and robot motion planning

Uniform models of algebraic computation.

  • Algebraic versions of the P=NP? question.
  • Its relationship to the classical question.
  • Complexity of the “Hilbert’s Nullstellensatz”.
  • Role of quantifier elimination.

Non-uniform models of algebraic computation.

  • Algebraic decision trees, branching programs, PRAM.
  • Lower bounds in these models.

Lecture notes

The lecture notes for the class are available either individually (below) or as one pdf file (1.3 M) .

Topics covered

  • Lecture 1: Overview; Preliminaries.
  • Lecture 2: Polynomials; The factorization and GCD problems.
  • Lecture 3: Polynomials and error-correcting codes. The Euclidean algorithm for GCD. Applications. Resultant.
  • Lecture 4: Properties of the resultant. Applications. Towards factorization of polynomials: Finding square roots modulo a prime.
  • Lecture 5: Factorization of univariate polynomials over finite fields.
  • Lecture 6: Berlekamp’s deterministic algorithm for factoring univariate polynomials. Existence, density and properties of irreducible polynomials.
  • Lecture 7: Factoring bivariate polynomials. Hensel’s lifting.
  • Lecture 8: Factoring bivariate polynomials (contd.). Digression: Applications to error-correction algorithms.
  • Lecture 9: Irreducibility testing; Black-box factoring of multivariate polynomials.
  • Lecture 10: Factoring polynomials over the rationals; Reduction to basis reduction.
  • Lecture 11: LLL’s Basis reduction algorithm.
  • Lecture 12: (2nd Phase of course) Ideals and Varieties. Division in Ideals; Groebner bases.
  • Lecture 13: Construction and uniqueness of Groebner bases. Solution to the Ideal Membership problem.
  • Lecture 14: Upper bound on the degrees for ideal generation. Complexity lower bound.
  • Lecture 15: Varieties. Emptiness of a variety. Elimination. Hilbert’s Nullstellensatz.
  • Lecture 16: Strong form of Hilbert’s Nullstellensatz. Quantifier Elimination.
  • Lecture 17: Quantifier elimination (contd.); Bezout’s Thm.
  • Lecture 18: Bezout’s Thm. and some applications.
  • Lecture 19: Algebraic models of computation; Ben-Or Cleve result.
  • Lecture 20: Blum-Shub-Smale model of computation (contd.) Undecidability of Mandelbrot Set.
  • Lecture 21: Algebaic settings for the P=NP question
  • Lecture 22: An Arthur-Merlin proof for the Hilbert’s Nullstellensatz.
  • Lecture 23: Non-uniform lower bounds; Linear independence; algebraic independence; Strassen’s degree bound.
  • Lecture 24: Ben-Or’s lower bounds based on no. of connected components; Survey of other methods: Volume; Euler characteristic and Betti numbers.
  • Lecture 25: Mulmuley’s Algebraic PRAM without bit operations. Lower bounds for LP and Max Flow.
  • Lecture 26: Summary. What we didn’t cover …

Style files for scribes: sample.tex and preamble.tex.