Codes, Metrics, and Applications

Codes, Metrics, and Applications

Codes, metrics, and applications Alexander Barg University of Maryland IITP RAS, Moscow ISIT 2016 Problems: § Distance distribution of codes § Duality of linear codes § Ordered linear codes and matroid invariants § Channel models matched to the ordered metrics; applications to polar codes, wiretap channels § Association schemes and bounds on the size of codes § Further extensions Acknowledgment: NSF grants CCF1217245 “Ordered metrics and their applications” CCF1422955, CCF0916919, CCF0807411 Talk summary Coding theory in a class of metric spaces: combinatorial and information-theoretic results and applications. § Channel models matched to the ordered metrics; applications to polar codes, wiretap channels § Association schemes and bounds on the size of codes § Further extensions Acknowledgment: NSF grants CCF1217245 “Ordered metrics and their applications” CCF1422955, CCF0916919, CCF0807411 Talk summary Coding theory in a class of metric spaces: combinatorial and information-theoretic results and applications. Problems: § Distance distribution of codes § Duality of linear codes § Ordered linear codes and matroid invariants Acknowledgment: NSF grants CCF1217245 “Ordered metrics and their applications” CCF1422955, CCF0916919, CCF0807411 Talk summary Coding theory in a class of metric spaces: combinatorial and information-theoretic results and applications. Problems: § Distance distribution of codes § Duality of linear codes § Ordered linear codes and matroid invariants § Channel models matched to the ordered metrics; applications to polar codes, wiretap channels § Association schemes and bounds on the size of codes § Further extensions Acknowledgment: NSF grants CCF1217245 “Ordered metrics and their applications” CCF1422955, CCF0916919, CCF0807411 Talk summary Coding theory in a class of metric spaces: combinatorial and information-theoretic results and applications. Problems: § Distance distribution of codes § Duality of linear codes § Ordered linear codes and matroid invariants § Channel models matched to the ordered metrics; applications to polar codes, wiretap channels § Association schemes and bounds on the size of codes § Further extensions Collaborators: Punarbasu Purkayastha, Woomyoung Park, Marcelo Firer, Maksim Skriganov, Min Ye, Talha Gulcu Talk summary Coding theory in a class of metric spaces: combinatorial and information-theoretic results and applications. Problems: § Distance distribution of codes § Duality of linear codes § Ordered linear codes and matroid invariants § Channel models matched to the ordered metrics; applications to polar codes, wiretap channels § Association schemes and bounds on the size of codes § Further extensions Acknowledgment: NSF grants CCF1217245 “Ordered metrics and their applications” CCF1422955, CCF0916919, CCF0807411 Outline § I. Brief recap: Linear codes, Weight distributions and duality, orthogonal arrays § II. Applications of the ordered distance § Wireless § Reed-Solomon codes § Approximation theory § III. Results on linear ordered codes § Shape distributions and bounds on codes § Duality of linear codes for poset metrics § Channel models § Polar codes C(dual) n C Dual code “ ty P Fq : px; yq “ 0 @x P u C(dual) (dual) Weight distribution of : Bi ; i “ 0; 1;:::; n Weight enumerators: ř ř n n´i i n (dual) n´i i BCpx; yq “ i“0 Bix y ; BC(dual) px; yq “ i“0 Bi x y I. Linear codes C n A linear code Ă Fq G; H generator and parity-check matrices Weight distribution Bi; i “ 0; 1;:::; n, where Bi “ 7tx P C : wpxq “ iu C(dual) n C Dual code “ ty P Fq : px; yq “ 0 @x P u C(dual) (dual) Weight distribution of : Bi ; i “ 0; 1;:::; n Weight enumerators: ř ř n n´i i n (dual) n´i i BCpx; yq “ i“0 Bix y ; BC(dual) px; yq “ i“0 Bi x y I. Linear codes C n A linear code Ă Fq G; H generator and parity-check matrices Weight distribution Bi; i “ 0; 1;:::; n, where Bi “ 7tx P C : wpxq “ iu Weight distributions are useful for analyzing structural properties of codes; probability of error under MAP or incomplete decoding I. Linear codes C n A linear code Ă Fq G; H generator and parity-check matrices Weight distribution Bi; i “ 0; 1;:::; n, where Bi “ 7tx P C : wpxq “ iu C(dual) n C Dual code “ ty P Fq : px; yq “ 0 @x P u C(dual) (dual) Weight distribution of : Bi ; i “ 0; 1;:::; n Weight enumerators: ř ř n n´i i n (dual) n´i i BCpx; yq “ i“0 Bix y ; BC(dual) px; yq “ i“0 Bi x y Approach via Fourier analysis: 1 ¸n B(dual) “ B K piq; j “ 0; 1;:::; n j |C| i j i“0 where ˆ ˙ˆ ˙ ¸i i n ´ i K piq “ p´iq` pq ´ 1qj´` j ` j ´ ` `“0 is a Krawtchouk polynomial Linear codes and duality The MacWilliams Theorem: 1 BCpx; yq “ BC(dual) px ` pq ´ 1qy; x ´ yq |C(dual)| Approach via Fourier analysis: 1 ¸n B(dual) “ B K piq; j “ 0; 1;:::; n j |C| i j i“0 where ˆ ˙ˆ ˙ ¸i i n ´ i K piq “ p´iq` pq ´ 1qj´` j ` j ´ ` `“0 is a Krawtchouk polynomial Linear codes and duality The MacWilliams Theorem: 1 BCpx; yq “ BC(dual) px ` pq ´ 1qy; x ´ yq |C(dual)| One of the basic facts in coding theory. Used for: § Classification of codes over various domains § Estimates of error probability § Bounds on the size of codes in terms of distance § Extensions used in sphere packing, optimality of lattices Linear codes and duality The MacWilliams Theorem: 1 BCpx; yq “ BC(dual) px ` pq ´ 1qy; x ´ yq |C(dual)| Approach via Fourier analysis: 1 ¸n B(dual) “ B K piq; j “ 0; 1;:::; n j |C| i j i“0 where ˆ ˙ˆ ˙ ¸i i n ´ i K piq “ p´iq` pq ´ 1qj´` j ` j ´ ` `“0 is a Krawtchouk polynomial § ´ ¯ k n´k qy x ´ y BCpx; yq “ px ´ yq y ZC ; (Greene 1976) x ´ y y § This connection extends to higher support weights (B ’97) (Wei ’91, Ozarow-Wyner ’84). § Codes and matroids: If the code is considered as an Fq-representation of a matroid M on the set t1; 2;:::; nu; then ZCpx; yq is the Whitney function of M Linear codes and duality The MacWilliams Theorem: § Linear algebraic approach: C Let A Ă t1; 2;:::; řnu; ρA “ rankpGpAqq; k “ dim k´ρA |A|´ρA Define ZCpx; yq “ AĂrns x y : Then ZCpx; yq “ ZC(dual) py; xq ZCpx; yq is the Whitney rank-nullity function of C § This connection extends to higher support weights (B ’97) (Wei ’91, Ozarow-Wyner ’84). § Codes and matroids: If the code is considered as an Fq-representation of a matroid M on the set t1; 2;:::; nu; then ZCpx; yq is the Whitney function of M Linear codes and duality The MacWilliams Theorem: § Linear algebraic approach: C Let A Ă t1; 2;:::; řnu; ρA “ rankpGpAqq; k “ dim k´ρA |A|´ρA Define ZCpx; yq “ AĂrns x y : Then ZCpx; yq “ ZC(dual) py; xq ZCpx; yq is the Whitney rank-nullity function of C § ´ ¯ k n´k qy x ´ y BCpx; yq “ px ´ yq y ZC ; (Greene 1976) x ´ y y § Codes and matroids: If the code is considered as an Fq-representation of a matroid M on the set t1; 2;:::; nu; then ZCpx; yq is the Whitney function of M Linear codes and duality The MacWilliams Theorem: § Linear algebraic approach: C Let A Ă t1; 2;:::; řnu; ρA “ rankpGpAqq; k “ dim k´ρA |A|´ρA Define ZCpx; yq “ AĂrns x y : Then ZCpx; yq “ ZC(dual) py; xq ZCpx; yq is the Whitney rank-nullity function of C § ´ ¯ k n´k qy x ´ y BCpx; yq “ px ´ yq y ZC ; (Greene 1976) x ´ y y § This connection extends to higher support weights (B ’97) (Wei ’91, Ozarow-Wyner ’84). Linear codes and duality The MacWilliams Theorem: § Linear algebraic approach: C Let A Ă t1; 2;:::; řnu; ρA “ rankpGpAqq; k “ dim k´ρA |A|´ρA Define ZCpx; yq “ AĂrns x y : Then ZCpx; yq “ ZC(dual) py; xq ZCpx; yq is the Whitney rank-nullity function of C § ´ ¯ k n´k qy x ´ y BCpx; yq “ px ´ yq y ZC ; (Greene 1976) x ´ y y § This connection extends to higher support weights (B ’97) (Wei ’91, Ozarow-Wyner ’84). § Codes and matroids: If the code is considered as an Fq-representation of a matroid M on the set t1; 2;:::; nu; then ZCpx; yq is the Whitney function of M 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 OAp8; 4; 1; 3q : 0 1 1 1 1 0 1 1 1 1 0 1 1 1 1 0 OAs form an example of designs in association schemes (Delsarte ’73) Orthogonal arrays Consider a code C, and suppose that C(dual) (dual) dp q “ t ` 1; i.e., Bi “ 0; i “ 1; 2;:::; t Then C is called an orthogonal array of strength t (C. R. Rao, 1946+) 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 OAp8; 4; 1; 3q : 0 1 1 1 1 0 1 1 1 1 0 1 1 1 1 0 OAs form an example of designs in association schemes (Delsarte ’73) Orthogonal arrays Consider a code C, and suppose that C(dual) (dual) dp q “ t ` 1; i.e., Bi “ 0; i “ 1; 2;:::; t Then C is called an orthogonal array of strength t (C. R. Rao, 1946+) 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 OAp8; 4; 1; 3q : 0 1 1 1 1 0 1 1 1 1 0 1 1 1 1 0 Orthogonal arrays Consider a code C, and suppose that C(dual) (dual) dp q “ t ` 1; i.e., Bi “ 0; i “ 1; 2;:::; t Then C is called an orthogonal array of strength t (C. R. Rao, 1946+) 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 OAp8; 4; 1; 3q : 0 1 1 1 1 0 1 1 1 1 0 1 1 1 1 0 OAs form an example of designs in association schemes (Delsarte ’73) § Hamming distance § Lee distance § Levenshtein (edit) distance § `1 distance; Kendall tau metric; Chebyshev (`8) distance § Subspace distance § Ordered metrics (Niederreiter ’92; Brualdi et al., ’95; Rosenbloom-Tsfasman, ’97) M.Deza and E.

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    114 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us