{"title":"An algorithm for symbolic-numeric sparse interpolation of multivariate polynomials whose degree bounds are unknown","authors":"Dai Numahata, Hiroshi Sekigawa","doi":"10.1145/3096730.3096734","DOIUrl":null,"url":null,"abstract":"We consider the problem of sparse interpolation of a multivariate black-box polynomial in floatingpoint arithmetic. More specifically, we assume that we are given a black-box polynomial <i>f</i> (<i>x</i><sub>1</sub>,...<i>x</i><sub>n</sub>) = Σ<sup><i>t</i></sup><sub><i>j</i>=1</sub> <i>c</i><sub><i>j</i></sub><i>x</i><sub>1</sub><sup><i>d</i><sub><i>j</i>, 1</sub></sup> ...<i>x</i><sub><i>n</i></sub><sup><i>d</i><sub><i>j</i>, n</sub></sup> ∈ C[<i>x</i><sub>1</sub>,...,<i>x</i><sub><i>n</i></sub>] (<i>c<sub>j</sub></i> ≠ 0)and the number of terms <i>t</i>, and that we can evaluate the value of <i>f</i> (<i>x<sup>1</sup>,...,x<sub>n</sub>)</i> at any point in C<sup><i>n</i></sup> in floating-point arithmetic. The problem is to find the coefficients <i>c<sub>1</sub></i>, ..., <i>c<sub>t</sub></i> and the exponents <i>d<sub>1,1,</sub>..., d<sub>t,n</sub></i>. We propose an efficient algorithm to solve the problem.","PeriodicalId":7093,"journal":{"name":"ACM Commun. Comput. Algebra","volume":"2 1","pages":"18-20"},"PeriodicalIF":0.0000,"publicationDate":"2017-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"ACM Commun. Comput. Algebra","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3096730.3096734","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
We consider the problem of sparse interpolation of a multivariate black-box polynomial in floatingpoint arithmetic. More specifically, we assume that we are given a black-box polynomial f (x1,...xn) = Σtj=1cjx1dj, 1 ...xndj, n ∈ C[x1,...,xn] (cj ≠ 0)and the number of terms t, and that we can evaluate the value of f (x1,...,xn) at any point in Cn in floating-point arithmetic. The problem is to find the coefficients c1, ..., ct and the exponents d1,1,..., dt,n. We propose an efficient algorithm to solve the problem.