Greedy Maximization of Functions with Bounded Curvature under Partition Matroid Constraints

Authors: Tobias Friedrich, Andreas Göbel, Frank Neumann, Francesco Quinzan, Ralf Rothenberger2272-2279

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments The maximum entropy sampling problem. In this set of experiments we study the following problem: Given a set of random variables, find the most informative subset of variables subject to a side constraint as in Problem (1). ... We consider the Berkley Earth climate dataset 2. ... Finding the maximum directed cut of a graph. In this set of experiments we study the performance of GREEDY for maximum directed cut in unweighted graphs. We compare these results with the optimal solutions, which we found via an Integer Linear Program solved with the state-of-the-art solver Gurobi (Gurobi Optimization 2018).
Researcher Affiliation Academia 1Chair of Algorithm Engineering, Hasso Plattner Institute, Potsdam, Germany 2Optimisation and Logistics, School of Computer Science, The University of Adelaide, Adelaide, Australia
Pseudocode Yes Algorithm 1: The GREEDY algorithm. input: a function f : V R 0; disjoint subsets B1, . . . , Bk V ; integers d1, . . . , dk s.t. 0 di |Bi| , i [k]; output: an approximate global maximum S of f s.t. |S Bi| di, i [k]; S ; while |S| Pk i=1 di do let ω V maximizing f(S {ω}) f(S) and s.t. |(S {ω}) Bi| di, i [k]; S S {ω}; return S;
Open Source Code Yes Our experimental code is available on github1. (footnote: 1https://github.com/Ralf Rothenberger/Greedy Max)
Open Datasets Yes We consider the Berkley Earth climate dataset 2. This dataset combines 1.6 billion temperature reports from 16 preexisting data archives, for over 39.000 unique stations. (footnote: 2http://berkeleyearth.org/data/) ... The experiments were conducted on 20 instances from Network Repository (Rossi and Ahmed 2016).
Dataset Splits No No explicit statements on training, validation, or test dataset splits (e.g., percentages, sample counts, or specific predefined splits) were found.
Hardware Specification No No specific hardware details (e.g., GPU/CPU models, processor types, memory amounts, or detailed computer specifications) used for running experiments are provided in the paper.
Software Dependencies Yes We compare these results with the optimal solutions, which we found via an Integer Linear Program solved with the state-of-the-art solver Gurobi (Gurobi Optimization 2018).
Experiment Setup No No specific experimental setup details, such as concrete hyperparameter values, training configurations, or system-level settings for models, are provided in the paper.