Optimized Scoring Systems: Toward Trust in Machine Learning for Healthcare and Criminal Justice

@article{Rudin2018OptimizedSS,
  title={Optimized Scoring Systems: Toward Trust in Machine Learning for Healthcare and Criminal Justice},
  author={Cynthia Rudin and Berk Ustun},
  journal={Interfaces},
  year={2018},
  volume={48},
  pages={449-466}
}
The authors developed and implemented transparent machine-learning models that call into question the use of black-box machine-learning models in healthcare and criminal justice applications. 
Please Stop Explaining Black Box Models for High Stakes Decisions
  • C. Rudin
  • Computer Science, Mathematics
  • ArXiv
  • 2018
TLDR
There is a way forward – it is to design models that are inherently interpretable, rather than trying to explain black box models, which is likely to perpetuate bad practices and can potentially cause catastrophic harm to society. Expand
Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead
  • C. Rudin
  • Computer Science
  • Nature Machine Intelligence
  • 2019
TLDR
This Perspective clarifies the chasm between explaining black boxes and using inherently interpretable models, outlines several key reasons why explainable black boxes should be avoided in high-stakes decisions, identifies challenges to interpretable machine learning, and provides several example applications whereinterpretable models could potentially replace black box models in criminal justice, healthcare and computer vision. Expand
Interpretable Machine Learning Classifiers for Brain Tumour Survival Prediction
TLDR
This paper uses a novel brain tumour dataset to compare two interpretable rule list models against popular machine learning approaches for brain tumours survival prediction and shows that the rule lists were only slightly outperformed by the black box models. Expand
Machine Learning Interpretability: A Survey on Methods and Metrics
TLDR
A review of the current state of the research field on machine learning interpretability while focusing on the societal impact and on the developed methods and metrics is provided. Expand
Mortality Prediction in ICU Patients Using Machine Learning Models
TLDR
Two mortality prediction models using the support vector machine (SVM) and linear discriminant analysis are presented and can be helpful for clinical experts for better decision making regarding the utilization of ICU allocations. Expand
Learning Optimized Risk Scores
TLDR
A new machine learning approach to learn risk scores that can fit risk scores in a way that scales linearly in the number of samples, provides a certificate of optimality, and obeys real-world constraints without parameter tuning or post-processing is presented. Expand
Machine learning for policing: a case study on arrests in Chile
ABSTRACT Police agencies expend considerable effort to anticipate future incidences of criminal behaviour. Since a large proportion of crimes are committed by a small group of individuals, preventiveExpand
Better Decisions for Children with “Big Data”: Can Algorithms Promote Fairness, Transparency and Parental Engagement?
Most countries operate procedures to safeguard children, including removal from parents in serious cases. In England, care applications and numbers have risen sharply, however, with wide variationsExpand
Health Equity in Artificial Intelligence and Primary Care Research: Protocol for a Scoping Review
TLDR
The extent to which AI systems in primary care examine the inherent bias toward or against vulnerable populations and appraise how these systems have mitigated the impact of such biases during their development is summarized. Expand
A framework for effective application of machine learning to microbiome-based classification problems
TLDR
A reusable open-source pipeline to train, validate, and interpret ML models that identify microbial biomarkers of disease and highlights the importance of choosing an ML approach based on the goal of the study, as the choice will inform expectations of performance and interpretability. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 117 REFERENCES
Is There Any Logic to Using Logit Finding the Right Tool for the Increasingly Important Job of Risk Prediction
A discussion of the relative merits of statistical techniques, specifically, random forests, stochastic gradient boosting, and logistic regression. Language: en
Interpretable classifiers using rules and Bayesian analysis: Building a better stroke prediction model
TLDR
A generative model called Bayesian Rule Lists is introduced that yields a posterior distribution over possible decision lists that employs a novel prior structure to encourage sparsity and has predictive accuracy on par with the current top algorithms for prediction in machine learning. Expand
Development and validation of the Emergency Department Assessment of Chest pain Score and 2 h accelerated diagnostic protocol
TLDR
A chest pain score and accelerated diagnostic protocol that could safely increase the proportion of patients suitable for early discharge and save time and costs is derived and validated. Expand
APACHE II: a severity of disease classification system.
TLDR
The form and validation results of APACHE II, a severity of disease classification system that uses a point score based upon initial values of 12 routine physiologic measurements, age, and previous health status, are presented. Expand
Optimized Risk Scores
TLDR
This paper forms a principled approach to learn risk scores that are fully optimized for feature selection, integer coefficients, and operational constraints, and presents a new cutting plane algorithm to efficiently recover its optimal solution. Expand
Big Data and Predictive Analytics: Recalibrating Expectations
With the routine use of electronic health records (EHRs) in hospitals, health systems, and physician practices, there has been rapid growth in the availability of health care data over the lastExpand
Very Simple Classification Rules Perform Well on Most Commonly Used Datasets
  • R. Holte
  • Computer Science
  • Machine Learning
  • 2004
TLDR
On most datasets studied, the best of very simple rules that classify examples on the basis of a single attribute is as accurate as the rules induced by the majority of machine learning systems. Expand
Statistical Procedures for Forecasting Criminal Behavior
There is a substantial and powerful literature in statistics and computer science clearly demonstrating that modern machine learning procedures can forecast more accurately than conventionalExpand
Which method predicts recidivism best?: a comparison of statistical, machine learning and data mining predictive models
Using criminal population conviction histories of recent offenders, prediction mod els are developed that predict three types of criminal recidivism: general recidivism, violent recidivism and sexualExpand
Interpretable Classification Models for Recidivism Prediction
TLDR
A new method known as SLIM (Supersparse Linear Integer Models) is used to produce accurate, transparent, and interpretable models along the full ROC curve, which can be used for decision-making for many different use cases, since they are just as accurate as the most powerful black-box machine learning models. Expand
...
1
2
3
4
5
...