Rating
1749
Battle Count: 50
Relevance
6/10
Highly relevant for risk management and fraud detection pipelines which are critical in fintech and trading infrastructure. While not directly about trading strategies, feature selection for risk models is a key component of quantitative finance operations.
Implementation Complexity
8/10
High complexity due to the integration of three distinct solver paradigms (classical ILP/QUBO, photonic entropic computing, and photonic boson sampling), requiring specific hardware access or specialized simulators and careful routing of methods to solvers.
Reproducibility
3/5
Datasets are public (ULB Kaggle, AmEx Kaggle). Solvers include commercial (Gurobi) and proprietary hardware/simulators (Dirac-3, Piquasso). Specific configurations and hyperparameters are detailed, but access to specific quantum hardware instances may be limited.
About this paper
Methodology: Solver-Aware Feature Selection Benchmark. Problem types: Classification, Feature Selection, Optimization, Imbalanced Learning, Risk Management.
The interactive Everscope explorer (charts, battles, favorites) loads below.