ABSTRACT
As demonstrated by media attention and research, Artificial Intelligence systems are not adequately addressing issues of fairness and bias, and more education on these topics is needed in industry and higher education. Currently, computer science courses that cover AI fairness and bias focus on statistical analysis or, on the other hand, attempt to bring in philosophical perspectives that lack actionable takeaways for students. Based on long-standing pedagogical research demonstrating the importance of using tools and visualizations to reinforce student learning, this case study reports on the impacts of using publicly-available visualization tools used in HCI practice as a resource for students examining algorithmic fairness concepts. Through qualitative review and observations of four focus groups, we examined six open-source fairness tools that enable students to visualize, quantify and explore algorithmic biases. The findings of this study provide insights into the benefits, challenges, and opportunities of integrating fairness tools as part of machine learning education.
Supplemental Material
- [1] 2021. https://www.kaggle.com/c/boston-housingGoogle Scholar
- [2] 2021. https://docs.responsibly.ai/notebooks/demo-fico-analysis.htmlGoogle Scholar
- [3] 2021. https://github.com/propublica/compas-analysis/Google Scholar
- 2021. Coded Bias.Google Scholar
- Alekh Agarwal, Alina Beygelzimer, Miroslav Dudík, John Langford, and Hanna Wallach. 2018. A reductions approach to fair classification. In International Conference on Machine Learning. PMLR, 60–69.Google Scholar
- Alekh Agarwal, Miroslav Dudik, and Zhiwei Steven Wu. 2019. Fair regression: Quantitative definitions and reduction-based algorithms. In International Conference on Machine Learning. PMLR, 120–129.Google Scholar
- Saleema Amershi, Giuseppe Carenini, Cristina Conati, Alan K Mackworth, and David Poole. 2008. Pedagogy and usability in interactive algorithm visualizations: Designing and evaluating CIspace. Interacting with Computers 20, 1 (2008), 64–96.Google ScholarDigital Library
- Hubert Baniecki, Wojciech Kretowicz, Piotr Piatyszek, Jakub Wisniewski, and Przemyslaw Biecek. 2020. dalex: Responsible Machine Learning with Interactive Explainability and Fairness in Python. arXiv preprint arXiv:2012.14406(2020).Google Scholar
- Niels Bantilan. 2017. scikit-ml. https://themis-ml.readthedocs.io/en/latest/Google Scholar
- Solon Barocas, Moritz Hardt, and Arvind Narayanan. 2019. Fairness and Machine Learning. fairmlbook.org. http://www.fairmlbook.org.Google Scholar
- Benjamin S Baumer, Randi L Garcia, Albert Y Kim, Katherine M Kinnaird, and Miles Q Ott. 2020. Integrating data science ethics into an undergraduate major. arXiv preprint arXiv:2001.07649(2020).Google Scholar
- Przemysław Biecek. 2018. DALEX: explainers for complex predictive models in R. The Journal of Machine Learning Research 19, 1 (2018), 3245–3249.Google ScholarDigital Library
- Huili Chen, Cheng Fu, Jishen Zhao, and Farinaz Koushanfar. 2019. DeepInspect: A Black-box Trojan Detection and Mitigation Framework for Deep Neural Networks.. In IJCAI. 4658–4664.Google Scholar
- Alexandra Chouldechova. 2017. Fair prediction with disparate impact: A study of bias in recidivism prediction instruments. Big data 5, 2 (2017), 153–163.Google Scholar
- Alexandra Chouldechova and Aaron Roth. 2020. A snapshot of the frontiers of fairness in machine learning. Commun. ACM 63, 5 (2020), 82–89.Google ScholarDigital Library
- Sam Corbett-Davies, Emma Pierson, Avi Feller, Sharad Goel, and Aziz Huq. 2017. Algorithmic decision making and the cost of fairness. In Proceedings of the 23rd acm sigkdd international conference on knowledge discovery and data mining. 797–806.Google ScholarDigital Library
- Kate Crawford. 2017. “The Trouble with Bias”.Google Scholar
- Michael Davis. 2006. Integrating ethics into technical courses: Micro-insertion. Science and Engineering Ethics 12, 4 (2006), 717–730.Google ScholarCross Ref
- William Dieterich, Christina Mendoza, and Tim Brennan. 2016. COMPAS risk scales: Demonstrating accuracy equity and predictive parity. Northpoint Inc 7, 7.4 (2016), 1.Google Scholar
- Dheeru Dua and Casey Graff. 2017. UCI Machine Learning Repository. http://archive.ics.uci.edu/mlGoogle Scholar
- Cynthia Dwork, Moritz Hardt, Toniann Pitassi, Omer Reingold, and Richard Zemel. 2012. Fairness through awareness. In Proceedings of the 3rd innovations in theoretical computer science conference. 214–226.Google ScholarDigital Library
- Facebook. 2020. FairFlow. https://www.facebook.com/FairFlowTech/Google Scholar
- Barbara J Grosz, David Gray Grant, Kate Vredenburgh, Jeff Behrends, Lily Hu, Alison Simmons, and Jim Waldo. 2019. Embedded EthiCS: integrating ethics across CS education. Commun. ACM 62, 8 (2019), 54–61.Google ScholarDigital Library
- Moritz Hardt, Eric Price, and Nathan Srebro. 2016. Equality of opportunity in supervised learning. arXiv preprint arXiv:1610.02413(2016).Google Scholar
- Shlomi Hod. 2018–. Responsibly: Toolkit for Auditing and Mitigating Bias and Fairness of Machine Learning Systems. http://docs.responsibly.ai/Google Scholar
- Anna Lauren Hoffmann and Katherine Alejandra Cross. 2021. Teaching Data Ethics: Foundations and Possibilities from Engineering and Computer Science Ethics Education. (2021).Google Scholar
- Michael Kearns and Aaron Roth. 2019. The ethical algorithm: The science of socially aware algorithm design. Oxford University Press.Google Scholar
- Jon Kleinberg, Sendhil Mullainathan, and Manish Raghavan. 2016. Inherent trade-offs in the fair determination of risk scores. arXiv preprint arXiv:1609.05807(2016).Google Scholar
- Ari Korhonen and Lauri Malmi. 2002. Matrix: concept animation and algorithm simulation system. In Proceedings of the Working Conference on Advanced Visual Interfaces. 109–114.Google ScholarDigital Library
- Michelle Seng Ah Lee and Jat Singh. 2021. The landscape and gaps in open source fairness toolkits. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems. ACM New York, NY, USA, 1–13.Google ScholarDigital Library
- A Markham. 2016. OKCupid data release fiasco: It’s time to rethink ethics education. Points: Data & Society(2016).Google Scholar
- Jacob Metcalf, Kate Crawford, and Emily F Keller. 2015. Pedagogical approaches to data ethics. Council for Big Data, Ethics, and Society(2015).Google Scholar
- Thomas L Naps, Guido Rößling, Vicki Almstrum, Wanda Dann, Rudolf Fleischer, Chris Hundhausen, Ari Korhonen, Lauri Malmi, Myles McNally, Susan Rodger, 2002. Exploring the role of visualization and engagement in computer science education. In Working group reports from ITiCSE on Innovation and technology in computer science education. 131–152.Google Scholar
- Željko Obrenović. 2012. Rethinking HCI education: teaching interactive computing concepts based on the experiential learning paradigm. interactions 19, 3 (2012), 66–70.Google ScholarDigital Library
- Cathy O’neil. 2016. Weapons of math destruction: How big data increases inequality and threatens democracy. Crown.Google Scholar
- Kris Powers, Paul Gross, Steve Cooper, Myles McNally, Kenneth J Goldman, Viera Proulx, and Martin Carlisle. 2006. Tools for teaching introductory programming: what works?. In Proceedings of the 37th SIGCSE technical symposium on Computer science education. 560–561.Google ScholarDigital Library
- Jay Quedado. 2021. Fairness Visualization Tools in ML Education. https://molochxte.github.io/ML-Fairness-Tools-in-Education/Google Scholar
- Google Brain Team. 2015. Tensorflow. https://themis-ml.readthedocs.io/en/latest/Google Scholar
- K Varshney. 2018. Introducing AI fairness 360. IBM Research blog, September 19 (2018).Google Scholar
- Matthijs Vincent. 2019. scikit-fairness. https://github.com/koaning/scikit-fairnessGoogle Scholar
- James Wexler, Mahima Pushkarna, Tolga Bolukbasi, Martin Wattenberg, Fernanda Viégas, and Jimbo Wilson. 2019. The what-if tool: Interactive probing of machine learning models. IEEE transactions on visualization and computer graphics 26, 1(2019), 56–65.Google Scholar
- Blake Woodworth, Suriya Gunasekar, Mesrob I Ohannessian, and Nathan Srebro. 2017. Learning non-discriminatory predictors. In Conference on Learning Theory. PMLR, 1920–1953.Google Scholar
- Matthew Zook, Solon Barocas, Danah Boyd, Kate Crawford, Emily Keller, Seeta Peña Gangadharan, Alyssa Goodman, Rachelle Hollander, Barbara A Koenig, Jacob Metcalf, 2017. Ten simple rules for responsible big data research.Google Scholar
Recommendations
Airtime Fairness for IEEE 802.11 Multirate Networks
Under a multi rate network scenario, the IEEE 802.11 DCF MAC fails to provide air-time fairness for all competing stations since the protocol is designed for ensuring max-min throughput fairness and the maximum achievable throughput by any station gets ...
Tools to Support Self-Regulated Learning in Online Environments: Literature Review
Lifelong Technology-Enhanced LearningAbstractSelf-regulated learning (SRL) skills are especially important in Massive Open Online Courses (MOOCs), where teacher guidance is scarce, and learners must engage in their learning process trying to succeed and achieve their learning goals. However,...
The multiple effects of combined tools in computer-based learning environments
Learners with high performance avoidance and perceived usability levels decreased the use of the tools.Learners with high critical thinking increased tool use.Only one of the two employed tools influenced performance positively.Time spent on tool seems ...
Comments