Berk Ustun

Expert
@ustunb

Assistant Professor @ UCSD focusing on safety and interpretability in AI/ML.

risk-slim. simple customizable risk scores in python

141

actionable-recourse. python tools to check recourse in linear classification

77

slim-python. simple customizable scoring systems in python

40

ctfdist. repair classifiers with disparate impact

10

slim-matlab. Optimized scoring systems for classification problems in MATLAB

8

dcptree. Code for the paper Fairness without Harm

7

concept-benchmark. Python

6

fairuse. Code for ICML 2023 Paper: "When Personalization Harms Performance: Reconsidering the use of Group Attributes in Prediction"

5

reachml. Python

5

classification-pipeline. R

4

recidivism-prediction. code for the paper: interpretable tools for recidivism prediction

4

spa. Selective Preference Aggregation

3

ldasearch. Python

3

cshock. Cardiogenic Shock

2

miplib2017-slim. SLIM instances for MIPLIB 2017

1

responsiveness. A Python library for recourse verification

1
16
Apply