Heterogeneous multitask learning with joint sparsity constraints

By Xiaolin Yang, Seyoung Kim, Eric P. Xing. 2009

92 citations recorded by the source corpus.

Citations in the corpus

• A Survey on Multi-Task Learning (2017, neutral): “[7], the heterogeneous MTL is considered to consist of different types of supervised tasks including classification and regression problems, and here we generalize it to a more general setting”
• Domain Generalization by Marginal Transfer Learning (2017, unknown): “Multi-task learning is similar to DG, except only the training tasks are of interest, and the goal is to leverage the similarity among distributions to improve the learning of individual predictors for each task (Caruana, 1997; Evgeniou et al., 2005; Yang et al., 2009).”
• Asynchronous Multi-task Learning (2016, unknown): “Note we assume that the tasks can be heterogeneous [9], i.e., some tasks can be regression while the other tasks are classification.”
• Privacy-Preserving Distributed Multi-Task Learning with Asynchronous Updates (2017, unknown): “Examples of commonly used MTL regularizations include shared feature subspace via low-rank modeling [2, 11] and shared feature subset via group sparsity [12, 31, 46].”
• $\ell_{p}-\ell_{q}$ Penalty for Sparse Linear and Sparse Multiple Kernel Multitask Learning (2011, unknown): “et al. and is motivated by applications in genetic association mapping [ 53 ].”
• Multitask learning and benchmarking with clinical time series data (2017, unknown): “Yang et al. [60] introduce a general framework for learning a mixture of regression and classification models with joint feature selection and apply it to gene expression data.”
• Flexible Clustered Multi-Task Learning by Learning Representative Tasks (2016, unknown): “This idea is called multi-task learning (MTL) [4], [10], [12], [23], [25], [34], [38], [41], [43], [51], [52], [58] and it attempts to share useful information across multiple relevant tasks by exploiting their intrinsic relationships.”
• Jointly Learning Heterogeneous Features for RGB-D Activity Recognition (2015, unknown): “For example, the work of [46] assumes that different tasks share a common set of input variables (i.”
• Robust Online Multi-Task Learning with Correlative and Personalized Structures (2017, unknown): “Tracenorm Regularized Multi-task Learning (TRML) [21] captured a common low-dimensional subspace of task relationship with a trace-norm regularization; or a common set of features [22], [23], e.”
• Multi-Modal Multi-Task Learning for Joint Prediction of Multiple Regression and Classification Variables in Alzheimer’s Disease (2011, unknown): “, 2008), and 2) sharing a common underlying representation acrossmultiple tasks (Argyriou et al., 2008; Obozinski et al., 2006; Yang et al., 2009).”
• Feature-based transfer learning with real-world applications (2010, unknown): “In Bioinformatics, motivated by that different biological entities, such as organisms, genes, etc, may be related to each other from a biological point of view, some research works have been proposed to apply transfer learning techniques to solve various computational biological problems, such as id”
• Toward an Architecture for Never-Ending Language Learning (2010, unknown): “Our work can also be seen as an example of multi-task learning in which several different functions are trained together , as in (Caruana 1997; Yang, Kim, and Xing 2009), in order to improve learning accuracy.”
• Robust multi-task feature learning (2012, unknown): “On the contrary, algorithms under implicit structure sharing do not explicitly impose all tasks to share certain parameters, but they implicitly capture some common structures; for example, the algorithms in [29, 24] constrain all tasks to share a common low rank subspace and the algorithms in [27, ”
• Integrating low-rank and group-sparse structures for robust multi-task learning (2011, unknown): “One key assumption in most of the existing MTL algorithms is that all tasks are correlated via a certain structure, which, for example, includes hidden units in neural networks [7], a common prior in a hierarchical Bayesian model [3, 29, 35, 37], parameters in Gaussian process covariance [17], kerne”
• Sparse multitask regression for identifying common mechanism of response to therapeutic targets (2010, unknown): “More recently, a heterogeneous multitask learning framework that considers both continuous (regression) and discrete (classification) variables was successfully used to discover genetic markers that jointly influence multiple correlated traits [33].”
• Efficient multi-task feature learning with calibration (2014, unknown): “…good regularizer to capture the underlying shared features among tasks; examples include group lasso multi-task feature learning [1, 19, 23, 25, 29, 33, 34], tree-guided group lasso multi-task feature learning [18], composite regularized multi-task feature learning [14, 16] and non-convex…”
• Heterogeneous Multi-task Learning for Human Pose Estimation with Deep Convolutional Neural Network (2014, unknown): “In [26], a heterogeneous multi-task model is trained by encouraging the parameters for the regression task and the classification task to share the same sparsity pattern.”