[Infolix]Fwd: [LABOLIX] Séminaire Robert Giegerich - Mercredi 16 Mai à 14h
Bonjour à tous,
Robert Giegerich, professeur à l'université technique de Bielefeld et invité DGAR de l'équipe AMIB pendant le mois de Mai, fera un exposé demain Mercredi 16 Mai à 14h en salle de séminaire du LIX. Ses thématiques de recherche sont très larges (allant de la compilation à la bioinformatique des ARN), avec un intérêt prononcé pour la programmation dynamique (algébrique) au sens large, et ses exposés toujours très agréables (peut être aurez vous le plaisir d'assister à une de ses chorégraphies destinées à comprendre le repliement des ARN...).
Vous y êtes donc les bienvenus, même si la bioinformatique n'est pas centrale à vos recherches.
Amicalement,
Yann
-------------------------------------------------------------------------------------------------------------------------------
Titre : Enjoy Dynamic Programming in Bellman's GAP
Orateur: Robert Giegerich
Applications of the dynamic programming paradigm are ubiquitous computer science, and especially so in bioinformatics. The discipline of algebraic dynamic programming liberates programmers from cumbersome coding and debugging, as algorithms can be described on a declarative level of abstraction. Teaching DP algorithms becomes more rewarding, as crucial ideas are better exhibited in a more abstract representation. The algebraic discipline underlies a good number of bioinformatics tools, such as RNAshapes, RNAlishapes, RNAhybrid, PknotsRG, KnotInFrame, Locomotif, and pKiss, some which are used widely.
Quite in contrast to this success, the proliferation of algebraic dynamic programming as a method remains marginal. It has been hindered by the fact that its original implementation was based on a Haskell-derived syntax and provided only marginal compiler optimizations.
The recent Bellman's GAP programming system provides a declarative language GAP-L with a Java-style syntax, designed to create dynamic programming algorithms over sequence data in algebraic style. The Bellman's GAP compiler generates code which is competitive to hand-coded DP recurrences, and arguably more reliable. It provides numerous features that would require considerable skills and eff ort in hand-coding, such as k-best analysis, full or stochastic backtracing, classified dynamic programming and algebra products, code generation for parallel hardware, and analysis of asymptotic space and runtime efficiency, including the consideration of space-time tradeo ffs. Several
bioinformatics tools have been re-implemented in Bellman's GAP by a mere change of syntax, with a considerable gain in efficiency.
If you have been a hard-coding dynamic programmer, try Bellman's GAP. It may bring joy into your life.
Bellman's GAP is joint work with Georg Sautho ff.
--
Robert Giegerich, professeur à l'université technique de Bielefeld et invité DGAR de l'équipe AMIB pendant le mois de Mai, fera un exposé demain Mercredi 16 Mai à 14h en salle de séminaire du LIX. Ses thématiques de recherche sont très larges (allant de la compilation à la bioinformatique des ARN), avec un intérêt prononcé pour la programmation dynamique (algébrique) au sens large, et ses exposés toujours très agréables (peut être aurez vous le plaisir d'assister à une de ses chorégraphies destinées à comprendre le repliement des ARN...).
Vous y êtes donc les bienvenus, même si la bioinformatique n'est pas centrale à vos recherches.
Amicalement,
Yann
-------------------------------------------------------------------------------------------------------------------------------
Titre : Enjoy Dynamic Programming in Bellman's GAP
Orateur: Robert Giegerich
Applications of the dynamic programming paradigm are ubiquitous computer science, and especially so in bioinformatics. The discipline of algebraic dynamic programming liberates programmers from cumbersome coding and debugging, as algorithms can be described on a declarative level of abstraction. Teaching DP algorithms becomes more rewarding, as crucial ideas are better exhibited in a more abstract representation. The algebraic discipline underlies a good number of bioinformatics tools, such as RNAshapes, RNAlishapes, RNAhybrid, PknotsRG, KnotInFrame, Locomotif, and pKiss, some which are used widely.
Quite in contrast to this success, the proliferation of algebraic dynamic programming as a method remains marginal. It has been hindered by the fact that its original implementation was based on a Haskell-derived syntax and provided only marginal compiler optimizations.
The recent Bellman's GAP programming system provides a declarative language GAP-L with a Java-style syntax, designed to create dynamic programming algorithms over sequence data in algebraic style. The Bellman's GAP compiler generates code which is competitive to hand-coded DP recurrences, and arguably more reliable. It provides numerous features that would require considerable skills and eff ort in hand-coding, such as k-best analysis, full or stochastic backtracing, classified dynamic programming and algebra products, code generation for parallel hardware, and analysis of asymptotic space and runtime efficiency, including the consideration of space-time tradeo ffs. Several
bioinformatics tools have been re-implemented in Bellman's GAP by a mere change of syntax, with a considerable gain in efficiency.
If you have been a hard-coding dynamic programmer, try Bellman's GAP. It may bring joy into your life.
Bellman's GAP is joint work with Georg Sautho ff.
--