{"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/correlational-neural-networks","title":"Correlational Neural Networks","arxiv_id":"1504.07225","date":"2015-04-27","proceeding":null,"authors":["Sarath Chandar","Mitesh M. Khapra","Hugo Larochelle","Balaraman Ravindran"],"abstract":"Common Representation Learning (CRL), wherein different descriptions (or\nviews) of the data are embedded in a common subspace, is receiving a lot of\nattention recently. Two popular paradigms here are Canonical Correlation\nAnalysis (CCA) based approaches and Autoencoder (AE) based approaches. CCA\nbased approaches learn a joint representation by maximizing correlation of the\nviews when projected to the common subspace. AE based methods learn a common\nrepresentation by minimizing the error of reconstructing the two views. Each of\nthese approaches has its own advantages and disadvantages. For example, while\nCCA based approaches outperform AE based approaches for the task of transfer\nlearning, they are not as scalable as the latter. In this work we propose an AE\nbased approach called Correlational Neural Network (CorrNet), that explicitly\nmaximizes correlation among the views when projected to the common subspace.\nThrough a series of experiments, we demonstrate that the proposed CorrNet is\nbetter than the above mentioned approaches with respect to its ability to learn\ncorrelated common representations. Further, we employ CorrNet for several cross\nlanguage tasks and show that the representations learned using CorrNet perform\nbetter than the ones learned using other state of the art approaches.","url_abs":"http://arxiv.org/abs/1504.07225v3","url_pdf":"http://arxiv.org/pdf/1504.07225v3.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":"correlational-neural-networks","repo_url":"https://github.com/apsarath/CorrNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"correlational-neural-networks","repo_url":"https://github.com/georgepar/synesthesia","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"}],"methods":[{"method_slug":"ae","method_name":"AE"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1504.07225","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}