ann 905
Abstract for An Investigation of Practical Approximate Nearest Neighbor Algorithms - Semantic Scholar
25 days ago by foodbaby
This paper concerns approximate nearest neighbor searching algorithms, which have become increasingly important, especially in high dimensional perception areas such as computer vision, with dozens of publications in recent years. Much of this enthusiasm is due to a successful new approximate nearest neighbor approach called Locality Sensitive Hashing (LSH). In this paper we ask the question: can earlier spatial data structure approaches to exact nearest neighbor, such as metric trees, be altered to provide approximate answers to proximity queries and if so, how? We introduce a new kind of metric tree that allows overlap: certain datapoints may appear in both the children of a parent. We also introduce new approximate k-NN search algorithms on this structure. We show why these structures should be able to exploit the same randomprojection-based approximations that LSH enjoys, but with a simpler algorithm and perhaps with greater efficiency. We then provide a detailed empirical evaluation on five large, high dimensional datasets which show up to 31-fold accelerations over LSH. This result holds true throughout the spectrum of approximation levels.
ANN
LSH
random-projection
metric
tree
25 days ago by foodbaby
Vectors in Search – Towards More Semantic Matching - Simon Hughes, Di…
9 weeks ago by foodbaby
Approximate Nearest Neighbor Search • Faster than full k-NN, with some loss in accuracy • Approaches can be either: • Data Dependent • Learns and adjusts from the data • Makes indexing new documents hard • Data Independent • Some Approaches: • KD Tree • LSH • Heuristic Methods • K-Means Tree • Randomized KD Forest • Paper: https://arxiv.org/abs/1603.09596 • HNSW (Hierarchical Navigable Small World Graphs – Top on http://ann-benchmarks.com/ • Paper: https://arxiv.org/pdf/1603.09320.pdf • Vector Thresholding • Choice of similarity metric is important in choosing an algorithm
ANN
vectors
search
9 weeks ago by foodbaby
The Unreasonable Effectiveness of Recurrent Neural Networks
9 weeks ago by ids
Amazing article showing the generation of new things from trained RNNs. Includes one of my favourite examples.- machine generated Shakespeare.
rnn
recurrentneuralnetwork
ann
artificialneuralnetwork
ai
machinelearning
deeplearning
9 weeks ago by ids
related tags
'diamond' 'stitch' 18th 1995-12-15 2001-12-07 2017 2018 a actually adam adkins adventures advice affleck after ai alan alfonso algo algorithm algorithms allison alvin and andrea andy angie anime anime_webpages anna anneorshine annoy anthony anxiety appeals applebaum ar arbor ardant artificial-neural-network artificial artificialneuralnetwork asai ashley augmented aunt aws ayako azure bag baker balm barnthouse barry becker becky behalikova ben ben_leslie benchmark benchmarks benn beringer bernie bigdata bill bird birt birthday bob bom bonnie book booker boone borgman bow brad brandt brown browning bruel bruno budge bujo bul... bullet burleigh but butter c++ c caan cake campagna carl carlton carmen carol carrera casey cecelia chaplin charles chase cheadle chives christie christine christopher cline clooney cluster-network clustering cnn collection colon combs committing compagnie computer conkle connelly content contractors convolutional cool cooper coulter course courses coworking coyle cpp crenna crotty curatola d damon dana danielle database datascience dataset david de deal: deep-learning deep deep_learning deeplearning dees del delano deliver? dempster denis depression derby derek designers detection dickinson discuss disney dl doc does doll don dori dowd: duality due dunn duran ed eddie edinburgh education elizabeth elliott elsa emanuel enterprise eric europe eva excellent eydie fake fanny fans fever fighting fiore fischbacher fletcher florence fondant for ford forster-delmas framework francois frank frankie franz free friendship from fromreese fullerton galinski gallagher gallienne gantt garcia gary gaudio gay genty george giamatti gnocchi go gomez google-cloud-platform google gorme got gould grace graphics greg gregory griffin guide guillaume gustavo halford handmade harrison hartmann hashing hd hearts hearts: helena henry hernandez herrera hibernian hibs history hit holly holmes hope horn how-to how howto howtocookthat humphrey ia ideas ifttt image important in index ines instrument integrated intelligence interpretability ir ira is italian ivey j jackson jacob jaime james javascript jb jeb!’ jemison jennifer jensen jeremy jerry jesse jim jo-jo joanna joe john johnson jolie jonathan jorge joshua joss journal joy jp jr julia june kain katia kawahara kd-tree kelly kenneth kentaro keras kinnear kish klitschko knn kordick kozel kristina kumlin kuster l la lampley language larry latisha laura lauren laurence laurent law lawless lawrence le learning leckie leeann leitch lemercier lennon lennox lesson lessons lewis library linderholm lithuania lloyd lori lotter lowe lsh lum luneau-lipton lydia lyndonville m mac macdonald machine-learning machine machinelearning make maker mann manoux mantella marchand margo marie mark martin martindale mary matsuo matt matthew mckernan mclane me medium merchant metcalf metric meyers michael micheline michigan miguel miklos miriam mit ml models mome mondal mooc mourning mycole nail nancy nearest-neighbor-search nearest-neighbor nearest-neighbors nearestneighbors nee need neil nek neopaganism network networks neural-network neural neuralnetworks news newsfeed newton nlp nmslib nolan nygard object occult october oears of on online optical optics optimization orlin ormond oruche overview paganism painting parameter parks parksandrecreation parrulli patrick patton paul pepper perez peter peters phil philippa philippe philosophy phina pike pinterest pitt plan playing pleshaw polish politics post306 pq presley princ... princess programming psychology python qin quin quinn r raising random-projection randy ranking ranknet read reality reardon recipe recognition recurrentneuralnetwork reed reese ref reggi reiner rep. repair research retrieval rhodes richard rick rl rnn roads&kingdoms robert roberts robin robinson robotham rockley roehm ronald ronn rosales rose rosenthal roy rpa russa rusty sachs sarah sastre schwaar schwartz schwary science scott sean search seqgan serps set setup shane shaobo sharing shutdown siching siegfried silva similarity simone slams smith-cameron snay soderbergh soeda software—and son song songbirds sontag space spark speech spotify sr sri steele stefano stenson steve steven steward strategy susan sweat tartini technology teen tensorflow the thelema therapy thomas threads tim to toderi tommy tool topher traffic tree troubles troy trump tutorial unified up ussr uutiset v valerie van var vector vectors velde vennema vermont video vincent vision vitale voted w wagner wall. wants ward watched watercolor watson wayne weintraub west wheeler william winning witchcraft with wittgenstein wladimir wood woodrow wordpress yahoo yann yuman ‘we 公主蛋糕教程Copy this bookmark: