The paper by Anshumali Shrivastava and Ping Li on “Asymmetric LSH (ALSH) for Sub-linear Time Maximum Inner Product Search (MIPS)” has won the “NIPS 2014 Best Paper Award”. The MIPS problem has widespread applications in search, learning, and optimization. This paper developed a simple “trick” which, by applying asymmetric transformations on the data and query separately, converts the MIPS problem into a conventional near neighbor search problem for which efficient sublinear time algorithms exist. This is the second best paper award Shrivastava and Li won in 2014. The other paper is about “A New Space for Comparing Graphs” which won the award in ASONAM 2014.
