{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/srp-efficient-class-aware-embedding-learning","title":"SRP: Efficient class-aware embedding learning for large-scale data via supervised random projections","arxiv_id":"1811.03166","date":"2018-11-07","proceeding":null,"authors":["Amir-Hossein Karimi","Alexander Wong","Ali Ghodsi"],"abstract":"Supervised dimensionality reduction strategies have been of great interest.\nHowever, current supervised dimensionality reduction approaches are difficult\nto scale for situations characterized by large datasets given the high\ncomputational complexities associated with such methods. While stochastic\napproximation strategies have been explored for unsupervised dimensionality\nreduction to tackle this challenge, such approaches are not well-suited for\naccelerating computational speed for supervised dimensionality reduction.\nMotivated to tackle this challenge, in this study we explore a novel direction\nof directly learning optimal class-aware embeddings in a supervised manner via\nthe notion of supervised random projections (SRP). The key idea behind SRP is\nthat, rather than performing spectral decomposition (or approximations thereof)\nwhich are computationally prohibitive for large-scale data, we instead perform\na direct decomposition by leveraging kernel approximation theory and the\nsymmetry of the Hilbert-Schmidt Independence Criterion (HSIC) measure of\ndependence between the embedded data and the labels. Experimental results on\nfive different synthetic and real-world datasets demonstrate that the proposed\nSRP strategy for class-aware embedding learning can be very promising in\nproducing embeddings that are highly competitive with existing supervised\ndimensionality reduction methods (e.g., SPCA and KSPCA) while achieving 1-2\norders of magnitude better computational performance. As such, such an\nefficient approach to learning embeddings for dimensionality reduction can be a\npowerful tool for large-scale data analysis and visualization.","url_abs":"http://arxiv.org/abs/1811.03166v1","url_pdf":"http://arxiv.org/pdf/1811.03166v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"srp-efficient-class-aware-embedding-learning","repo_url":"https://github.com/lightonai/supervised-random-projections","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"dimensionality-reduction","task_name":"Dimensionality Reduction"},{"task_slug":"supervised-dimensionality-reduction","task_name":"Supervised dimensionality reduction"}],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}