KDD paper-573 preview: Explaining the Predictions of Any Classifier

Understanding the reasons behind predictions is an important requirement for using them in decision-making. This work introduces an explanation technique, named LIME, that spells out the predictions of any classifier in an interpretable and faithful manner. It does so by learning an interpretable model locally around the prediction. Watch the three minute video — HERE

Additional videos, highlighting papers presented at this year’s event can be found at our YouTube Channel — https://www.youtube.com/channel/UCPsUUDUlcTJuP-fRa7z85aQ

Remember that KDD takes place August 13–17, 2016 at the Hilton SF, Union Square: https://www.regonline.com/Register/Checkin.aspx?EventID=1832521