Graphical Multi-Task Learning

By D. Sheldon. 2008

43 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Kernels for Vector-Valued Functions: a Review (2011, methodology): “Following [64, 91], we can define a regularizer that, in addition to a standard regularization on the single components, forces stronger or weaker similarity between them through a givenD×D positive weight matrix M,”
Vector Field Learning via Spectral Filtering (2010, background): “In this section, we briefly review some matrix valued kernels for vector fields learning - see [25,26,27].”
Multi-Task Averaging (2011, methodology): “lti-task learning, and Evgeniou et al. (2005) showed that the problem of estimating Ttask functions with certain types of regularization can be cast as a single-task problem using multi-task kernels. Sheldon (2008) built on this work by proposing a graphical multi-task learning framework using graph”
Multi-output learning via spectral filtering (2012, methodology): “The material in this section is largely drawn from previous results of Evgeniou et al. (2005), Sheldon (2008), Caponnetto et al. (2008).”
Input Output Kernel Regression (2016, background): “The matrix A 2 was proposed by Evgeniou et al. (2005) and Sheldon (2008) for multi-task learning from the following regularizer: This regularization term forces two tasks h and h to be close to each other when the similarity value M is high and conversely.”
Revisiting Stein's paradox: multi-task averaging (2014, background): “The most closely related work is that of Sheldon (2008) and Kato et al. (2008), where the regularizer or constraint, respectively, is T∑ r=1 T∑ s=1 Ars‖βr − βs‖22, which is the MTA regularizer if applied to mean estimation.”
Domain adaptation from multiple sources via auxiliary classifiers (2009, methodology): “2007; Sheldon, 2008 ) and can be reduced to a form, which is similar to the dual of †-SVR: minfi;fi⁄ 1 2 (fiifi⁄)0K(fiifi⁄)+y0(fiifi⁄)+†10(fi+fi⁄)”
Predictive Learning in the Presence of Heterogeneity and Limited Training Data (2014, background): “For instance, the similarity among related tasks can be represented in the form of a graph which can then be used for regularizing the learning over each individual task [6, 13].”