A MaxSAT-Based Framework for Group Testing

Authors: Lorenzo Ciampiconi, Bishwamittra Ghosh, Jonathan Scarlett, Kuldeep S Meel10144-10152

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our extensive experimental results show that MGT can solve group testing instances of 10000 items with 3% defectivity, which no prior work can handle to the best of our knowledge.
Researcher Affiliation Academia Lorenzo Ciampiconi,1 Bishwamittra Ghosh,2 Jonathan Scarlett,2 Kuldeep S Meel2 1Politecnico di Milano, 2National University of Singapore
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code Yes 1https://github.com/meelgroup/mgt
Open Datasets No We consider Bernoulli trials (Aldridge, Johnson, and Scarlett 2019) to model a stochastic group testing instance. In particular, we consider three independent and identically distributed (i.i.d.) Bernoulli processes: an item is defective independently with probability p < 0.5, item i belongs to test j independently with probability q, and in the noisy setting, a test outcome is inverted independently with probability d < 0.5. The paper generates its own data according to these processes rather than using a publicly available dataset with concrete access information.
Dataset Splits No We repeat each experiment for l = 100 trials to ensure statistical consistency. The paper does not use standard train/validation/test splits, as it generates synthetic problem instances rather than using a fixed dataset.
Hardware Specification Yes The experiment was conducted on a machine with Intel core i7 (3.4 GHz) and 8 GB of RAM.
Software Dependencies No In our implementation, we employ Max HS (Davies and Bacchus 2011) as the underlying Max SAT solver. [...] In the LP approach, we use CPLEX as the underlying LP solver3. Version numbers for the solvers are not provided.
Experiment Setup Yes We set the cut-off time of both the LP and Max HS solvers to be 100 seconds. [...] In particular, we consider three independent and identically distributed (i.i.d.) Bernoulli processes: an item is defective independently with probability p < 0.5, item i belongs to test j independently with probability q, and in the noisy setting, a test outcome is inverted independently with probability d < 0.5.