Explainable Machine Learning for Real World Resource Constrained Problems

Machine Learning (ML) has contributed to many advances in science and technology, recently the applications in high-stake decision-making has been increased. The ability to build statistically complex relationships between data and outcomes, without explicitly programming the rules makes it popular among the computerized decision-making community. The state-of-the-art models which performer well, are more complex and inexplicable. On the contrary, in high-stakes settings as healthcare, finance, and criminal justice, have strict ethical concerns which lead to explain each decision or the model in whole. Moreover, there it is necessary to ensure the fairness of the decisions and the transparency to avoid wrongful rationality. Furthermore the acts like General Data Protection Regulation (GDPR) and Personal Data Protection Act (PDPA) makes it mandatory to explain the decisions made based on statistical relationships and gave a right for users to demand an explanation.