Skip to main content

Table 5 Suggested solutions to discrimination in Big Data

From: Big Data and discrimination: perils, promises and solutions. A systematic review

Suggested solutions

Paper references

1. Computer science and technical solutions

 1.1. Pre-processing

Kamiran and Calders 2012 [42], Hajian and Domingo-Ferrer 2013 [33], Kamiran et al. 2013 [43], Hajian et al. 2014 [32]

 1.2. In-processing

Calders and Verwer 2010 [17], Pope and Sydnor 2011 [66], Kamiran et al. 2013 [43], Zliobaite and Custers 2016 [95], Kroll et al. 2017 [45]

 1.3. Post-processing

Hajian et al. 2015 [34]

 1.4.Mixed methods

d'Alessandro et al. 2017 [25]

 1.5. Implementation of transparency

Hildebrandt and Koops 2010 [35], Schermer 2011 [73], Citron and Pasquale 2014 [21], Kroll et al. 2017 [45]

 1.6. Privacy preserving strategies

Hildebrandt and Koops 2010 [35], Hajian et al. 2015 [34]

 1.7. Exploratory fairness analysis

Veale and Binns 2017 [84]

2. Legal solutions

Hildebrandt and Koops 2010 [35], Hoffman 2010 [37], Citron and Pasquale 2014 [21], Peppet 2014 [62], Hirsch 2015 [36], Kuempel 2016 [46], Hoffman 2017 [38]

3. Human based solutions

 3.1. Human in the loop

Zarsky 2014 [93], Berendt and Preibusch 2017 [11], d'Alessandro et al. 2017 [25]

 3.2. Third parties

Mantelero 2016 [54], Veale and Binns 2017 [84]

 3.3. Multidisciplinary involvement

Cohen et al. 2014 [22], Taylor 2016 [77, 78], Taylor 2017 [79]

 3.4. Education

Zarsky 2014 [93], Veale and Binns 2017 [84]

 3.5. Implementing EHR flexibility

Hoffman 2010 [37]