Fair classification via Monte Carlo policy gradient method
Само за регистроване кориснике
2021
Чланак у часопису (Објављена верзија)
Метаподаци
Приказ свих података о документуАпстракт
Artificial intelligence is steadily increasing its impact on everyday life. Therefore, the societal issues of artificial intelligence have become an important concern in the AI research. The presence of data that reflects human biases towards historically discriminated groups defined by sensitive features such as race and gender, results in machine learning models which discriminate against these groups. In order to tackle the impact of bias in data, researchers developed a variety of specialized machine learning algorithms which are able to satisfy different fairness constraints imposed on the model. Group fairness constraints do not fit standard machine learning formulations easily due to their non-differentiable nature. In this paper we developed a technique for learning a fair classifier by Monte Carlo policy gradient method which naturally deals with such non-differentiable constraints. Our methodology focuses on direct optimization of both group fairness metric and predictive per...formance of the model. In addition, we propose two different variance reduction techniques of gradient estimation. We compare our models to seven other related and state-of-the-art models and demonstrate that they are able to achieve better trade-off between accuracy and unfairness. To the best of our knowledge, this is the first fair classification algorithm which solves the issue of non-differentiable constraints by reinforcement learning techniques.
Кључне речи:
Reinforcement learning / REINFORCE / Fairness / Deep learning / Combinatorial optimizationИзвор:
Engineering Applications of Artificial Intelligence, 2021, 104Издавач:
- Pergamon-Elsevier Science Ltd, Oxford
Финансирање / пројекти:
- ONR/ONR Global [N629091912008]
- company Saga New Frontier Group Belgrade
- Аутоматско резоновање и истраживање података (RS-MESTD-Basic Research (BR or ON)-174021)
- Иновативни приступ у примени интелигентних технолошких система за производњу делова од лима заснован на еколошким принципима (RS-MESTD-Technological Development (TD or TR)-35004)
- Интеракција етиопатогенетских механизама пародонтопатије и паериимплантитиса са системским болестима данашњице (RS-MESTD-Integrated and Interdisciplinary Research (IIR or III)-41008)
DOI: 10.1016/j.engappai.2021.104398
ISSN: 0952-1976
WoS: 000686249600009
Scopus: 2-s2.0-85111926472
Институција/група
Fakultet organizacionih naukaTY - JOUR AU - Petrović, Andrija AU - Nikolić, Mladen AU - Jovanović, Miloš AU - Bijanić, Miloš AU - Delibašić, Boris PY - 2021 UR - https://rfos.fon.bg.ac.rs/handle/123456789/2153 AB - Artificial intelligence is steadily increasing its impact on everyday life. Therefore, the societal issues of artificial intelligence have become an important concern in the AI research. The presence of data that reflects human biases towards historically discriminated groups defined by sensitive features such as race and gender, results in machine learning models which discriminate against these groups. In order to tackle the impact of bias in data, researchers developed a variety of specialized machine learning algorithms which are able to satisfy different fairness constraints imposed on the model. Group fairness constraints do not fit standard machine learning formulations easily due to their non-differentiable nature. In this paper we developed a technique for learning a fair classifier by Monte Carlo policy gradient method which naturally deals with such non-differentiable constraints. Our methodology focuses on direct optimization of both group fairness metric and predictive performance of the model. In addition, we propose two different variance reduction techniques of gradient estimation. We compare our models to seven other related and state-of-the-art models and demonstrate that they are able to achieve better trade-off between accuracy and unfairness. To the best of our knowledge, this is the first fair classification algorithm which solves the issue of non-differentiable constraints by reinforcement learning techniques. PB - Pergamon-Elsevier Science Ltd, Oxford T2 - Engineering Applications of Artificial Intelligence T1 - Fair classification via Monte Carlo policy gradient method VL - 104 DO - 10.1016/j.engappai.2021.104398 UR - conv_2544 ER -
@article{ author = "Petrović, Andrija and Nikolić, Mladen and Jovanović, Miloš and Bijanić, Miloš and Delibašić, Boris", year = "2021", abstract = "Artificial intelligence is steadily increasing its impact on everyday life. Therefore, the societal issues of artificial intelligence have become an important concern in the AI research. The presence of data that reflects human biases towards historically discriminated groups defined by sensitive features such as race and gender, results in machine learning models which discriminate against these groups. In order to tackle the impact of bias in data, researchers developed a variety of specialized machine learning algorithms which are able to satisfy different fairness constraints imposed on the model. Group fairness constraints do not fit standard machine learning formulations easily due to their non-differentiable nature. In this paper we developed a technique for learning a fair classifier by Monte Carlo policy gradient method which naturally deals with such non-differentiable constraints. Our methodology focuses on direct optimization of both group fairness metric and predictive performance of the model. In addition, we propose two different variance reduction techniques of gradient estimation. We compare our models to seven other related and state-of-the-art models and demonstrate that they are able to achieve better trade-off between accuracy and unfairness. To the best of our knowledge, this is the first fair classification algorithm which solves the issue of non-differentiable constraints by reinforcement learning techniques.", publisher = "Pergamon-Elsevier Science Ltd, Oxford", journal = "Engineering Applications of Artificial Intelligence", title = "Fair classification via Monte Carlo policy gradient method", volume = "104", doi = "10.1016/j.engappai.2021.104398", url = "conv_2544" }
Petrović, A., Nikolić, M., Jovanović, M., Bijanić, M.,& Delibašić, B.. (2021). Fair classification via Monte Carlo policy gradient method. in Engineering Applications of Artificial Intelligence Pergamon-Elsevier Science Ltd, Oxford., 104. https://doi.org/10.1016/j.engappai.2021.104398 conv_2544
Petrović A, Nikolić M, Jovanović M, Bijanić M, Delibašić B. Fair classification via Monte Carlo policy gradient method. in Engineering Applications of Artificial Intelligence. 2021;104. doi:10.1016/j.engappai.2021.104398 conv_2544 .
Petrović, Andrija, Nikolić, Mladen, Jovanović, Miloš, Bijanić, Miloš, Delibašić, Boris, "Fair classification via Monte Carlo policy gradient method" in Engineering Applications of Artificial Intelligence, 104 (2021), https://doi.org/10.1016/j.engappai.2021.104398 ., conv_2544 .