CombOptNet: Fit the Right NP-Hard Problem by Learning Integer Programming Constraints

Logo poskytovatele

Varování

Publikace nespadá pod Ekonomicko-správní fakultu, ale pod Fakultu informatiky. Oficiální stránka publikace je na webu muni.cz.
Autoři

PAULUS Anselm ROLÍNEK Michal MUSIL Vít AMOS Brandon MARTIUS Georg

Rok publikování 2021
Druh Článek ve sborníku
Konference INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 139
Fakulta / Pracoviště MU

Fakulta informatiky

Citace
www https://proceedings.mlr.press/v139/paulus21a.html
Klíčová slova machine learning; combinatorial optimization; integer linear programming; constraints
Přiložené soubory
Popis Bridging logical and algorithmic reasoning with modern machine learning techniques is a fundamental challenge with potentially transformative impact. On the algorithmic side, many NP-HARD problems can be expressed as integer programs, in which the constraints play the role of their "combinatorial specification." In this work, we aim to integrate integer programming solvers into neural network architectures as layers capable of learning both the cost terms and the constraints. The resulting end-to-end trainable architectures jointly extract features from raw data and solve a suitable (learned) combinatorial problem with state-of-the-art integer programming solvers. We demonstrate the potential of such layers with an extensive performance analysis on synthetic data and with a demonstration on a competitive computer vision keypoint matching benchmark.
Související projekty:

Používáte starou verzi internetového prohlížeče. Doporučujeme aktualizovat Váš prohlížeč na nejnovější verzi.