ABSTRACT
The fairness characteristic is a critical attribute of trusted AI systems. A plethora of research has proposed diverse methods for individual fairness testing. However, they are suffering from three major limitations, i.e., low efficiency, low effectiveness, and model-specificity. This work proposes ExpGA, an explanation-guided fairness testing approach through a genetic algorithm (GA). ExpGA employs the explanation results generated by interpretable methods to collect high-quality initial seeds, which are prone to derive discriminatory samples by slightly modifying feature values. ExpGA then adopts GA to search discriminatory sample candidates by optimizing a fitness value. Benefiting from this combination of explanation results and GA, ExpGA is both efficient and effective to detect discriminatory individuals. Moreover, ExpGA only requires prediction probabilities of the tested model, resulting in a better generalization capability to various models. Experiments on multiple real-world benchmarks, including tabular and text datasets, show that ExpGA presents higher efficiency and effectiveness than four state-of-the-art approaches.
- 2018. AEQUITAS Github Project. https://github.com/sakshiudeshi/Aequitas.Google Scholar
- 2020. ADF Github Project. https://github.com/pxzhang94/ADF.Google Scholar
- 2021. Bank Marketing Dataset. https://archive.ics.uci.edu/ml/datasets/Bank+Marketing.Google Scholar
- 2021. Census Income Dataset. http://archive.ics.uci.edu/ml/datasets/Census+Income.Google Scholar
- 2021. ConceptNet: An open, multilingual knowledge graph. https://conceptnet.io/.Google Scholar
- 2021. Fairness 360: Understand and mitigate bias in ML models. https://ai-fairness-360.org/.Google Scholar
- 2021. German Credit Dataset. https://archive.ics.uci.edu/ml/datasets/statlog+(german+credit+data).Google Scholar
- 2021. IMDB Dataset. https://www.imdb.com/interfaces/.Google Scholar
- 2021. ML-fairness-gym: A Tool for Exploring Long-Term Impacts of Machine Learning Systems. https://ai.googleblog.com/2020/02/ml-faimess-gym-tool-for-exploring-long.html.Google Scholar
- 2021. Principle (a): Lawfulness, fairness and transparency. https://ico.org.uk/for-organisations/guide-to-data-protection/guide-to-the-general-data-protection-regulation-gdpr/principles/lawfulness-fairness-and-transparency/.Google Scholar
- 2021. The Role of Protected Attributes in AI Fairness. https://www.trustscience.com/blog/the-role-of-protected-attributes-in-ai-fairness.Google Scholar
- 2021. Scikit-Learn. https://scikit-learn.org/stable/user_guide.html.Google Scholar
- 2021. Stanford Sentiment Treebank Dataset. http://nlpprogress.com/english/sentiment_analysis.html.Google Scholar
- 2021. TensorFlow. https://github.com/tensorflow/tensorflow.Google Scholar
- Aniya Aggarwal, Pranay Lohia, Seema Nagar, Kuntal Dey, and Diptikalyan Saha. 2019. Black box fairness testing of machine learning models. In Proc. FSE. 625--635.Google ScholarDigital Library
- Andrea Arcuri and Lionel Briand. 2011. A practical guide for using statistical tests to assess randomized algorithms in software engineering. In Proc. ICSE. IEEE, 1--10.Google ScholarDigital Library
- Yuriy Brun and Alexandra Meliou. 2018. Software fairness. In Proc. FSE. 754--759.Google ScholarDigital Library
- Joy Buolamwini and Timnit Gebru. 2018. Gender shades: Intersectional accuracy disparities in commercial gender classification. In Proc. FAT. 77--91.Google Scholar
- Joymallya Chakraborty, Suvodeep Majumder, and Tim Menzies. 2021. Bias in Machine Learning Software: Why? How? What to do? arXiv preprint arXiv:2105.12195 (2021).Google Scholar
- Sam Corbett-Davies and Sharad Goel. 2018. The measure and mismeasure of fairness: A critical review of fair machine learning. arXiv preprint arXiv:1808.00023 (2018).Google Scholar
- Brian d'Alessandro, Cathy O'Neil, and Tom LaGatta. 2017. Conscientious classification: A data scientist's guide to discrimination-aware classification. Big data 5, 2 (2017), 120--134.Google Scholar
- Ming Fan, Wenying Wei, Xiaofei Xie, Yang Liu, Xiaohong Guan, and Ting Liu. 2020. Can We Trust Your Explanations? Sanity Checks for Interpreters in Android Malware Analysis. IEEE Transactions on Information Forensics and Security 16 (2020), 838--853.Google ScholarCross Ref
- James R Foulds, Rashidul Islam, Kamrun Naher Keya, and Shimei Pan. 2020. An intersectional definition of fairness. In Proc. ICDE. 1918--1921.Google ScholarCross Ref
- Sainyam Galhotra, Yuriy Brun, and Alexandra Meliou. 2017. Fairness testing: testing software for discrimination. In Proc. FSE. 498--510.Google ScholarDigital Library
- Riccardo Guidotti, Anna Monreale, Salvatore Ruggieri, Dino Pedreschi, Franco Turini, and Fosca Giannotti. 2018. Local rule-based explanations of black box decision systems. arXiv preprint arXiv:1805.10820 (2018).Google Scholar
- Wenbo Guo, Dongliang Mu, Jun Xu, Purui Su, Gang Wang, and Xinyu Xing. 2018. Lemna: Explaining deep learning based security applications. In Proc. CCS. 364--379.Google ScholarDigital Library
- Scott M Lundberg and Su-In Lee. 2017. A unified approach to interpreting model predictions. In Proc. NeurIPS. 4765--4774.Google Scholar
- Pingchuan Ma, Shuai Wang, and Jin Liu. [n.d.]. Metamorphic testing and certified mitigation of fairness violations in nlp models. In Proc. IJCAI. 458--465.Google Scholar
- Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan. 2019. A survey on bias and fairness in machine learning. arXiv preprint arXiv:1908.09635 (2019).Google Scholar
- Deepti Bala Mishra, Rajashree Mishra, Arup Abhinna Acharya, and Kedar Nath Das. 2019. Test data generation for mutation testing using genetic algorithm. In Soft Computing for Problem Solving. 857--867.Google Scholar
- Shira Mitchell, Eric Potash, Solon Barocas, Alexander D'Amour, and Kristian Lum. 2018. Prediction-based decisions and fairness: A catalogue of choices, assumptions, and definitions. arXiv preprint arXiv:1811.07867 (2018).Google Scholar
- Dino Pedreshi, Salvatore Ruggieri, and Franco Turini. [n.d.]. Discrimination-aware data mining. In Proc. KDD. 560--568.Google Scholar
- Jeffrey Pennington, Richard Socher, and Christopher D Manning. 2014. Glove: Global vectors for word representation. In Proc. EMNLP. 1532--1543.Google ScholarCross Ref
- Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. 2016. why should i trust you?: Explaining the predictions of any classifier. In Proc. KDD. 1135--1144.Google ScholarDigital Library
- Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. 2018. Anchors: High-precision model-agnostic explanations. In Proc. AAAI. 1527--1535.Google ScholarCross Ref
- Nripsuta Ani Saxena. 2019. Perceptions of Fairness. In Proc. AIES. 537--538.Google ScholarDigital Library
- Akshat Sharma, Rishon Patani, and Ashish Aggarwal. 2016. Software testing using genetic algorithms. International Journal of Computer Science & Engineering Survey 7, 2 (2016), 21--33.Google ScholarCross Ref
- Rachael Tatman. 2017. Gender and dialect bias in YouTube automatic captions. In Proc. EthNLP@EACL. 53--59.Google Scholar
- Florian Tramer, Vaggelis Atlidakis, Roxana Geambasu, Daniel Hsu, Jean-Pierre Hubaux, Mathias Humbert, Ari Juels, and Huang Lin. 2017. Fairtest: Discovering unwarranted associations in data-driven applications. In Proc. EuroS&P. 401--416.Google ScholarCross Ref
- Sakshi Udeshi, Pryanshu Arora, and Sudipta Chattopadhyay. 2018. Automated directed fairness testing. In Proc. ASE. 98--108.Google ScholarDigital Library
- Alexander Warnecke, Daniel Arp, Christian Wressnegger, and Konrad Rieck. 2020. Evaluating explanation methods for deep learning in security. In Proc. EuroS&P. 158--174.Google ScholarCross Ref
- Peixin Zhang, Jingyi Wang, Jun Sun, Guoliang Dong, Xinyu Wang, Xingen Wang, Jin Song Dong, and Ting Dai. 2020. White-box fairness testing through adversarial sampling. In Proc. ICSE. 949--960.Google ScholarDigital Library
- Ridong Zhang and Jili Tao. 2017. A nonlinear fuzzy neural network modeling approach using an improved genetic algorithm. IEEE Transactions on Industrial Electronics 65, 7 (2017), 5882--5892.Google ScholarCross Ref
Index Terms
- Explanation-guided fairness testing through genetic algorithm
Recommendations
An improved genetic algorithm with conditional genetic operators and its application to set-covering problem
The genetic algorithm (GA) is a popular, biologically inspired optimization method. However, in the GA there is no rule of thumb to design the GA operators and select GA parameters. Instead, trial-and-error has to be applied. In this paper we present an ...
Hybrid Taguchi-genetic algorithm for global numerical optimization
In this paper, a hybrid Taguchi-genetic algorithm (HTGA) is proposed to solve global numerical optimization problems with continuous variables. The HTGA combines the traditional genetic algorithm (TGA), which has a powerful global exploration capability,...
Solving Japanese nonograms by Taguchi-based genetic algorithm
A Taguchi-based genetic algorithm (TBGA) is proposed to solve Japanese nonogram puzzles. The TBGA exploits the power of global exploration inherent in the traditional genetic algorithm (GA) and the abilities of the Taguchi method in efficiently ...
Comments