{"type":"ScholarlyArticle","title":"Certifai: A Toolkit for Building Trust in AI Systems","authors":[{"name":"Jette Henderson","slug":"jette-henderson","given":"Jette","family":"Henderson","affiliation":""},{"name":"Shubham Sharma","slug":"shubham-sharma","given":"Shubham","family":"Sharma","affiliation":"The University of Texas , Austin"},{"name":"Alan Gee","slug":"alan-gee","given":"Alan","family":"Gee","affiliation":"The University of Texas , Austin"},{"name":"Valeri Alexiev","slug":"valeri-alexiev","given":"Valeri","family":"Alexiev","affiliation":""},{"name":"Steve Draper","slug":"steve-draper","given":"Steve","family":"Draper","affiliation":""},{"name":"Carlos Marin","slug":"carlos-marin","given":"Carlos","family":"Marin","affiliation":""},{"name":"Yessel Hinojosa","slug":"yessel-hinojosa","given":"Yessel","family":"Hinojosa","affiliation":""},{"name":"Christine Draper","slug":"christine-draper","given":"Christine","family":"Draper","affiliation":""},{"name":"Michael Perng","slug":"michael-perng","given":"Michael","family":"Perng","affiliation":""},{"name":"Luis Aguirre","slug":"luis-aguirre","given":"Luis","family":"Aguirre","affiliation":""},{"name":"Michael Li","slug":"michael-li","given":"Michael","family":"Li","affiliation":"The University of Texas , Austin"},{"name":"Sara Rouhani","slug":"sara-rouhani","given":"Sara","family":"Rouhani","affiliation":"The University of Texas , Dallas"},{"name":"Shorya Consul","slug":"shorya-consul","given":"Shorya","family":"Consul","affiliation":"The University of Texas , Austin"},{"name":"Susan Michalski","slug":"susan-michalski","given":"Susan","family":"Michalski","affiliation":""},{"name":"Akarsh Prasad","slug":"akarsh-prasad","given":"Akarsh","family":"Prasad","affiliation":""},{"name":"Mayank Chutani","slug":"mayank-chutani","given":"Mayank","family":"Chutani","affiliation":""},{"name":"Aditya Kumar","slug":"aditya-kumar","given":"Aditya","family":"Kumar","affiliation":""},{"name":"Shahzad Alam","slug":"shahzad-alam","given":"Shahzad","family":"Alam","affiliation":""},{"name":"Prajna Kandarpa","slug":"prajna-kandarpa","given":"Prajna","family":"Kandarpa","affiliation":""},{"name":"Binnu Jesudasan","slug":"binnu-jesudasan","given":"Binnu","family":"Jesudasan","affiliation":""},{"name":"Colton Lee","slug":"colton-lee","given":"Colton","family":"Lee","affiliation":""},{"name":"Michael Criscolo","slug":"michael-criscolo","given":"Michael","family":"Criscolo","affiliation":""},{"name":"Sinead Williamson","slug":"sinead-williamson","given":"Sinead","family":"Williamson","affiliation":"The University of Texas , Austin"},{"name":"Matt Sanchez","slug":"matt-sanchez","given":"Matt","family":"Sanchez","affiliation":""},{"name":"Joydeep Ghosh","slug":"joydeep-ghosh","given":"Joydeep","family":"Ghosh","affiliation":"The University of Texas , Austin"}],"abstract":"As more companies and governments build and use machine learning models to automate decisions, there is an ever-growing need to monitor and evaluate these models' behavior once they are deployed. \r\nOur team at CognitiveScale has developed a toolkit called Cortex Certifai to answer this need. Cortex Certifai is a framework that assesses aspects of robustness, fairness, and interpretability of any classification or regression model trained on tabular data, without requiring access to its internal workings. Additionally, Cortex Certifai allows users to compare models along these different axes and only requires 1) query access to the model and 2) an “evaluation” dataset. At its foundation, Cortex Certifai generates counterfactual explanations, which are synthetic data points close to input data points but differing in terms of model prediction. The tool then harnesses characteristics of these counterfactual explanations to analyze different aspects of the supplied model and delivers evaluations relevant to a variety of different stakeholders (e.g., model developers, risk analysts, compliance officers). Cortex Certifai can be configured and executed using a command-line interface (CLI), within jupyter notebooks, or on the cloud, and the results are recorded in JSON files and can be visualized in an interactive console. Using these reports, stakeholders can understand, monitor, and build trust in their AI systems. In this paper, we provide a brief overview of a demonstration of Cortex Certifai's capabilities.","doi":"10.24963/ijcai.2020/759","url":"https://ijcai.org/proceedings/2020/759","pdf":"https://ijcai.org/proceedings/2020/0759.pdf","markdown":"https://ijcai.org/proceedings/2020/759.md","bibtex":"https://ijcai.org/proceedings/2020/bibtex/759","volume":{"slug":"2020","title":"Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2020","url":"https://ijcai.org/proceedings/2020/","year":2020,"location":"Yokohama","dates":"January 2021 (postponed from 11-17 July 2020)","editors":"Christian Bessiere","isbn":"978-0-9992411-6-5","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2020"},"section":"Demos","track":"Demos","number":759,"pages":{"first":5249,"last":5251},"keywords":["Machine Learning: general"],"videos":[],"references":[{"position":1,"raw":"[Mothilal et al., 2020] Ramaravind K Mothilal, Amit Sharma, and Chenhao Tan. Explaining machine learning classifiers through diverse counterfactual explanations. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, pages 607-617, 2020.","title":"Explaining machine learning classifiers through diverse counterfactual explanations","authors":"Mothilal","year":"2020","venue":"Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency","doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Pearl, 2000] Judea Pearl. Causality: models, reasoning and inference, volume 29. Springer, 2000.","title":"Judea Pearl. Causality: models, reasoning and inference","authors":"Pearl","year":"2000","venue":null,"doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Russell, 2019] Chris Russell. Efficient search for diverse coherent explanations. In Proceedings of the Conference on Fairness, Accountability, and Transparency, pages 10- 19. ACM, 2019.","title":"Efficient search for diverse coherent explanations","authors":"Chris Russell","year":"2019","venue":"Proceedings of the Conference on Fairness, Accountability, and Transparency","doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Sharma et al., 2020] Shubham Sharma, Jette Henderson, and Joydeep Ghosh. CERTIFAI: A common framework to provide explanations and analyse the fairness and robust- ness of black-box models. In 3rd AAAI/ACM Conference on AI, Ethics, and Society, pages 166-172, 2020.","title":"CERTIFAI: A common framework to provide explanations and analyse the fairness and robustness of black-box models","authors":"Shubham Sharma, Jette Henderson, Joydeep Ghosh","year":"2020","venue":"3rd AAAI/ACM Conference on AI, Ethics, and Society","doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Ustun et al., 2019] Berk Ustun, Alexander Spangher, and Yang Liu. Actionable recourse in linear classification. In Proceedings of the Conference on Fairness, Accountabil- ity, and Transparency, pages 10-19. ACM, 2019.","title":"Actionable recourse in linear classification","authors":"Berk Ustun, Alexander Spangher, Yang Liu","year":"2019","venue":"Proceedings of the Conference on Fairness, Accountability, and Transparency","doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Wachter et al., 2017] Sandra Wachter, Brent Mittelstadt, and Chris Russell. Counterfactual explanations without opening the black box: automated decisions and the gdpr. Harvard Journal of Law & Technology, 31(2), 2017.","title":"Counterfactual explanations without opening the black box: automated decisions and the gdpr","authors":"Sandra Wachter, Brent Mittelstadt, Chris Russell","year":"2017","venue":"Harvard Journal of Law & Technology","doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}