Large Scale Ranking
WNCG Prof. Joydeep Ghosh and student Sreangsu Acharyya are developing ranking algorithms that retain the simplicity of pointwise methods but have the full power of listwise methods. Our algorithms exploit the fact that for ranking, only the order of scores matters, not the actual score values. This flexibility has been exploited to obtain algorithms that scale well to very large datasets, and that provide increased accuracy and lower computational time than current state-of-the-art methods.
This research funded by the Schlumberger Chair.
Paper 1: MEMR: A Margin Equipped Monotone Retargeting Framework for Ranking (to appear)