Exploit Every Bit: Effective Caching for High-Dimensional Nearest Neighbor Search PROJECT TITLE :Exploit Every Bit: Effective Caching for High-Dimensional Nearest Neighbor SearchABSTRACT:High-dimensional nearest neighbor (kNN) search contains a wide range of applications in multimedia information retrieval. Existing disk-based mostly NN search methods incur important I/O prices within the candidate refinement section. In this paper, we tend to propose to cache compact approximate representations of knowledge points in main memory so as to scale back the candidate refinement time during NN search. This drawback raises 2 challenging issues: (i) that is the foremost effective encoding scheme for information points to support NN search? and (ii) what's the optimal range of bits for encoding a data purpose? For (i), we have a tendency to formulate and solve a novel histogram optimization problem that decides the most effective encoding theme. For (ii), we tend to develop a cost model for automatically tuning the optimal variety of bits for encoding points. Additionally, our approach is generic and applicable to exact / approximate NN search methods. Intensive experimental results on real datasets demonstrate that our proposal can accelerate the candidate refinement time of NN search by at least an order of magnitude. Did you like this research project? To get this research project Guidelines, Training and Code... Click Here facebook twitter google+ linkedin stumble pinterest Reflecting on Software Engineering Research for Internet Computing Bayesian $M$ -Ary Hypothesis Testing: The Meta-Converse and Verdú-Han Bounds Are Tight
