Abstract | ||
---|---|---|
Dimensionality reduction involves mapping a set of high dimensional input points onto a low dimensional manifold so that 'similar" points in input space are mapped to nearby points on the manifold. We present a method - called Dimensionality Reduction by Learning an Invariant Mapping (DrLIM) - for learning a globally coherent nonlinear function that maps the data evenly to the output manifold. The learning relies solely on neighborhood relationships and does not require any distancemeasure in the input space. The method can learn mappings that are invariant to certain transformations of the inputs, as is demonstrated with a number of experiments. Comparisons are made to other techniques, in particular LLE. |
Year | DOI | Venue |
---|---|---|
2006 | 10.1109/CVPR.2006.100 | Computer Vision and Pattern Recognition, 2006 IEEE Computer Society Conference |
Keywords | Field | DocType |
low dimensional manifold,coherent nonlinear function,dimensionality reduction,input space,neighborhood relationship,output manifold,nearby point,certain transformation,high dimensional input point,invariant mapping,feature extraction,geoscience,astronomy,image analysis,biology,data visualization,distance metric,manufacturing industries | Computer vision,Data visualization,Image generation,Nonlinear system,Dimensionality reduction,Pattern recognition,Computer science,Feature extraction,Artificial intelligence,Invariant (mathematics),Diffusion map,Manifold | Conference |
Volume | ISSN | ISBN |
2 | 1063-6919 | 0-7695-2597-0 |
Citations | PageRank | References |
488 | 19.70 | 12 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
R. Hadsell | 1 | 1678 | 100.80 |
Sumit Chopra | 2 | 2835 | 181.37 |
Yann LeCun | 3 | 26090 | 3771.21 |