RESUMO
Discrimination-aware classification methods remedy socioeconomic disparities exacerbated by machine learning systems. In this paper, we propose a novel data pre-processing technique that assigns weights to training instances in order to reduce discrimination without changing any of the inputs or labels. While the existing reweighing approach only looks into sensitive attributes, we refine the weights by utilizing both sensitive and insensitive ones. We formulate our weight assignment as a linear programming problem. The weights can be directly used in any classification model into which they are incorporated. We demonstrate three advantages of our approach on synthetic and benchmark datasets. First, discrimination reduction comes at a small cost in accuracy. Second, our method is more scalable than most other pre-processing methods. Third, the trade-off between fairness and accuracy can be explicitly monitored by model users. Code is available at https://github.com/frnliang/refined_reweighing.
Assuntos
Aprendizado de Máquina , Algoritmos , HumanosRESUMO
Chronic heart failure (CHF) and diabetes mellitus are associated with morbidity and mortality. CHF and diabetes generally simultaneously occur, resulting in adverse outcomes. Diabetes complicates cardiomyopathy and exacerbates heart failure conditions. An increase in natriuretic peptides, including atrial natriuretic peptide (ANP), and another endsogenously generated peptide, brain natriuretic peptide (BNP), serves an essential role in CHF. The aim of this study was to explore the molecular regulation between bone morphogenetic protein-2 (BMP-2) and ANP or BNP in diabetes-associated cardiomyopathy. In total, 25 serum samples were collected from patients with CHF with or without type 2 diabetes mellitus to compare with 25 controls. Cardiomyopathy and hyperglycemia were induced in rats by doxorubicin and streptozotocin, respectively. AC16 cells were used to study molecular mechanisms. BMP, ANP and BNP concentration in patients and rats were measured by ELISA. Flow cytometry was performed to analyze cell pyroptosis and ROS production. Reverse transcription-quantitative PCR and western blotting were used to examine mRNA and protein expression of NOD-, LRR- and pyrin domain-containing protein 3 (NLRP3), pro-caspase-1, caspase-1 (p20) and gasdermin D. BMP-2 was negatively correlated with ANP and BNP in CHF patients with type 2 diabetes mellitus. Similar results were obtained in rats and AC16 cells. BMP-2 decreased the NLRP3 inflammasome activation and cell pyroptosis. The present study found evidence that the cardioprotective effects of BMP-2 act through ANP and BNP both in vivo and in vitro. BMP-2 inhibits inflammasome formation. The results suggested that BMP-2 may serve as a novel therapeutic target for the treatment of diabetic heart conditions.