{"conference":{"name":"NIPS SISO: NIPS Workshop on \"Structured Input - Structured Output\""},"publication_status":"published","publisher":"Curran Associates, Inc.","citation":{"apa":"Lampert, C., &#38; Blaschko, M. (2008). Joint kernel support estimation for structured prediction (pp. 1–4). Presented at the NIPS SISO: NIPS Workshop on “Structured Input - Structured Output,” Curran Associates, Inc.","short":"C. Lampert, M. Blaschko, in:, Curran Associates, Inc., 2008, pp. 1–4.","ieee":"C. Lampert and M. Blaschko, “Joint kernel support estimation for structured prediction,” presented at the NIPS SISO: NIPS Workshop on “Structured Input - Structured Output,” 2008, pp. 1–4.","chicago":"Lampert, Christoph, and Matthew Blaschko. “Joint Kernel Support Estimation for Structured Prediction,” 1–4. Curran Associates, Inc., 2008.","mla":"Lampert, Christoph, and Matthew Blaschko. <i>Joint Kernel Support Estimation for Structured Prediction</i>. Curran Associates, Inc., 2008, pp. 1–4.","ama":"Lampert C, Blaschko M. Joint kernel support estimation for structured prediction. In: Curran Associates, Inc.; 2008:1-4.","ista":"Lampert C, Blaschko M. 2008. Joint kernel support estimation for structured prediction. NIPS SISO: NIPS Workshop on ‘Structured Input - Structured Output’, 1–4."},"type":"conference","_id":"3706","title":"Joint kernel support estimation for structured prediction","month":"12","year":"2008","day":"12","quality_controlled":0,"date_published":"2008-12-12T00:00:00Z","date_updated":"2021-01-12T07:51:37Z","status":"public","date_created":"2018-12-11T12:04:43Z","abstract":[{"text":"We present a new technique for structured prediction that works in a hybrid generative/discriminative way, using a one-class support vector machine to model the joint probability of (input, output)-pairs in a joint reproducing kernel Hilbert space. Compared to discriminative techniques, like conditional random fields or structured output SVMs?, the proposed method has the advantage that its training time depends only on the number of training examples, not on the size of the label space. Due to its generative aspect, it is also very tolerant against ambiguous, incomplete or incorrect labels. Experiments on realistic data show that our method works efficiently and robustly in situations that discriminative techniques have problems with or that are computationally infeasible for them.","lang":"eng"}],"main_file_link":[{"open_access":"0","url":"http://agbs.kyb.tuebingen.mpg.de/wikis/bg/siso2008/Blaschkoetal.pdf"}],"page":"1 - 4","publist_id":"2650","author":[{"id":"40C20FD2-F248-11E8-B48F-1D18A9856A87","last_name":"Lampert","full_name":"Christoph Lampert","first_name":"Christoph","orcid":"0000-0001-8622-7887"},{"last_name":"Blaschko","first_name":"Matthew","full_name":"Blaschko,Matthew B"}],"extern":1}