Low-Rank Approximations of Hyperbolic Embeddings

Low-Rank Approximations of Hyperbolic Embeddings. Jawanpuria, P., Meghwanshi, M., & Mishra, B.

The hyperbolic manifold is a smooth manifold of negative constant curvature. While the hyperbolic manifold is well-studied in the literature, it has gained interest in the machine learning and natural language processing communities lately due to its usefulness in modeling continuous hierarchies. Tasks with hierarchical structures are ubiquitous in those fields and there is a general interest to learning hyperbolic representations or embeddings of such tasks. Additionally, these embeddings of related tasks may also share a low-rank subspace. In this work, we propose to learn hyperbolic embeddings such that they also lie in a low-dimensional subspace. In particular, we consider the problem of learning a low-rank factorization of hyperbolic embeddings. We cast these problems as manifold optimization problems and propose computationally efficient algorithms. Empirical results illustrate the efficacy of the proposed approach.

@article{jawanpuriaLowrankApproximationsHyperbolic2019,
  archivePrefix = {arXiv},
  eprinttype = {arxiv},
  eprint = {1903.07307},
  primaryClass = {cs, math, stat},
  title = {Low-Rank Approximations of Hyperbolic Embeddings},
  url = {http://arxiv.org/abs/1903.07307},
  abstract = {The hyperbolic manifold is a smooth manifold of negative constant curvature. While the hyperbolic manifold is well-studied in the literature, it has gained interest in the machine learning and natural language processing communities lately due to its usefulness in modeling continuous hierarchies. Tasks with hierarchical structures are ubiquitous in those fields and there is a general interest to learning hyperbolic representations or embeddings of such tasks. Additionally, these embeddings of related tasks may also share a low-rank subspace. In this work, we propose to learn hyperbolic embeddings such that they also lie in a low-dimensional subspace. In particular, we consider the problem of learning a low-rank factorization of hyperbolic embeddings. We cast these problems as manifold optimization problems and propose computationally efficient algorithms. Empirical results illustrate the efficacy of the proposed approach.},
  urldate = {2019-03-19},
  date = {2019-03-18},
  keywords = {Statistics - Machine Learning,Computer Science - Machine Learning,Mathematics - Optimization and Control},
  author = {Jawanpuria, Pratik and Meghwanshi, Mayank and Mishra, Bamdev},
  file = {/home/dimitri/Nextcloud/Zotero/storage/HSJ9EB3M/Jawanpuria et al. - 2019 - Low-rank approximations of hyperbolic embeddings.pdf;/home/dimitri/Nextcloud/Zotero/storage/88FXB97C/1903.html}
}

Downloads: 0

{"_id":"cSpM4p5Mx9zJSve93","bibbaseid":"jawanpuria-meghwanshi-mishra-lowrankapproximationsofhyperbolicembeddings","authorIDs":[],"author_short":["Jawanpuria, P.","Meghwanshi, M.","Mishra, B."],"bibdata":{"bibtype":"article","type":"article","archiveprefix":"arXiv","eprinttype":"arxiv","eprint":"1903.07307","primaryclass":"cs, math, stat","title":"Low-Rank Approximations of Hyperbolic Embeddings","url":"http://arxiv.org/abs/1903.07307","abstract":"The hyperbolic manifold is a smooth manifold of negative constant curvature. While the hyperbolic manifold is well-studied in the literature, it has gained interest in the machine learning and natural language processing communities lately due to its usefulness in modeling continuous hierarchies. Tasks with hierarchical structures are ubiquitous in those fields and there is a general interest to learning hyperbolic representations or embeddings of such tasks. Additionally, these embeddings of related tasks may also share a low-rank subspace. In this work, we propose to learn hyperbolic embeddings such that they also lie in a low-dimensional subspace. In particular, we consider the problem of learning a low-rank factorization of hyperbolic embeddings. We cast these problems as manifold optimization problems and propose computationally efficient algorithms. Empirical results illustrate the efficacy of the proposed approach.","urldate":"2019-03-19","date":"2019-03-18","keywords":"Statistics - Machine Learning,Computer Science - Machine Learning,Mathematics - Optimization and Control","author":[{"propositions":[],"lastnames":["Jawanpuria"],"firstnames":["Pratik"],"suffixes":[]},{"propositions":[],"lastnames":["Meghwanshi"],"firstnames":["Mayank"],"suffixes":[]},{"propositions":[],"lastnames":["Mishra"],"firstnames":["Bamdev"],"suffixes":[]}],"file":"/home/dimitri/Nextcloud/Zotero/storage/HSJ9EB3M/Jawanpuria et al. - 2019 - Low-rank approximations of hyperbolic embeddings.pdf;/home/dimitri/Nextcloud/Zotero/storage/88FXB97C/1903.html","bibtex":"@article{jawanpuriaLowrankApproximationsHyperbolic2019,\n archivePrefix = {arXiv},\n eprinttype = {arxiv},\n eprint = {1903.07307},\n primaryClass = {cs, math, stat},\n title = {Low-Rank Approximations of Hyperbolic Embeddings},\n url = {http://arxiv.org/abs/1903.07307},\n abstract = {The hyperbolic manifold is a smooth manifold of negative constant curvature. While the hyperbolic manifold is well-studied in the literature, it has gained interest in the machine learning and natural language processing communities lately due to its usefulness in modeling continuous hierarchies. Tasks with hierarchical structures are ubiquitous in those fields and there is a general interest to learning hyperbolic representations or embeddings of such tasks. Additionally, these embeddings of related tasks may also share a low-rank subspace. In this work, we propose to learn hyperbolic embeddings such that they also lie in a low-dimensional subspace. In particular, we consider the problem of learning a low-rank factorization of hyperbolic embeddings. We cast these problems as manifold optimization problems and propose computationally efficient algorithms. Empirical results illustrate the efficacy of the proposed approach.},\n urldate = {2019-03-19},\n date = {2019-03-18},\n keywords = {Statistics - Machine Learning,Computer Science - Machine Learning,Mathematics - Optimization and Control},\n author = {Jawanpuria, Pratik and Meghwanshi, Mayank and Mishra, Bamdev},\n file = {/home/dimitri/Nextcloud/Zotero/storage/HSJ9EB3M/Jawanpuria et al. - 2019 - Low-rank approximations of hyperbolic embeddings.pdf;/home/dimitri/Nextcloud/Zotero/storage/88FXB97C/1903.html}\n}\n\n","author_short":["Jawanpuria, P.","Meghwanshi, M.","Mishra, B."],"key":"jawanpuriaLowrankApproximationsHyperbolic2019","id":"jawanpuriaLowrankApproximationsHyperbolic2019","bibbaseid":"jawanpuria-meghwanshi-mishra-lowrankapproximationsofhyperbolicembeddings","role":"author","urls":{"Paper":"http://arxiv.org/abs/1903.07307"},"keyword":["Statistics - Machine Learning","Computer Science - Machine Learning","Mathematics - Optimization and Control"],"downloads":0},"bibtype":"article","biburl":"https://raw.githubusercontent.com/dlozeve/newblog/master/bib/all.bib","creationDate":"2020-01-08T20:39:39.305Z","downloads":0,"keywords":["statistics - machine learning","computer science - machine learning","mathematics - optimization and control"],"search_terms":["low","rank","approximations","hyperbolic","embeddings","jawanpuria","meghwanshi","mishra"],"title":"Low-Rank Approximations of Hyperbolic Embeddings","year":null,"dataSources":["3XqdvqRE7zuX4cm8m"]}