Multi-domain learning by confidence-weighted parameter combination

By Mark Dredze, Alex Kulesza, Koby Crammer. 2010

138 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems, Optimization & Training

How researchers see this paper

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

In Artificial Intelligence

In Information Systems

In Optimization & Training

Citations in the corpus

Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges (2019, background): “Many-to-one multilingual NMT can be categorized as a multi-domain 3 learning problem (Dredze et al., 2010; Joshi et al., 2012; Nam and Han, 2015), where the task of translating into a selected language remains the same, but the input distribution is different across source languages (Arivazhagan et…”
Learning Multi-domain Convolutional Neural Networks for Visual Tracking (2015, background): “, sentiment classification with multiple products and spam filtering with multiple users), and various approaches have been proposed [6, 7, 23].”
Deep Multi-task Representation Learning: A Tensor Factorisation Approach (2016, background): “Multi-domain learning [Dredze et al., 2010, Joshi et al., 2012] is a relatively independent line of research to multi-task learning.”
Portfolio Selection : A Survey (2014, unknown): “…to further exploit the second-order portfolio information, which refers to the variance of portfolio weights (not price or price relative), following the mean reversion trading idea via Confidence Weighted (CW) online learning [Dredze et al. 2008; Crammer et al. 2008, 2009; Dredze et al. 2010].”
A survey of multi-source domain adaptation (2015, background): “The most popular performance measurement in the classification task for multi-source domain adaptation is the classification accuracy [26,5,11,30,34].”
A Unified Perspective on Multi-Domain and Multi-Task Learning (2014, background): “Multi-Domain Learning Multi-Domain Learning (MDL) (Dredze et al., 2010; Joshi et al., 2012) shares properties of both domain adaptation and multi-task learning.”
Book Reviews: Data-Intensive Text Processing with MapReduce by Jimmy Lin and Chris Dyer (2010, background): “Experiments, however, show that the merged solution is inferior to the output of running the training algorithm on the entire dataset [52].”
Semi-Supervised Learning and Domain Adaptation in Natural Language Processing (2013, background): “Pointers to representative relevant work, both in NLP (Daumé III 2007; Finkel and Manning 2009) and machine learning (BenDavid et al. 2010; Dredze, Kulesza, and Crammer 2010), would also have been an asset.”
Adversarial Transfer Learning (2018, background): “Transfer learning is defined as the learning scenario where a model is trained on a source domain or task and evaluated on a different but related target domain or task, where either the tasks or domains (or both) differ [31, 96, 122].”
Multi-task Domain Adaptation for Sequence Tagging (2016, unknown): “…(i.e. linking similar features during learning) (Dredze and Cram-mer, 2008; Daum´e III, 2007, 2009; Finkel and Manning, 2009; Kumar et al., 2010; Dredze et al., 2010), and learning cross domain representations (Blitzer et al., 2006, 2007; Glorot et al., 2011; Chen et al., 2012; Yang and…”
A Survey of Unsupervised Deep Domain Adaptation (2018, methodology): “defined as the learning scenario where a model is trained on a source domain or task and evaluated on a different but related target domain or task, where either the tasks or domains (or both) differ [61, 80, 180, 252]. For instance, we may wish to learn a model on a handwritten digit dataset (e.g.,”
Multi-Domain Adversarial Learning (2019, background): “In MDL, the different domains can be taken into account by maintaining shared and domain-specific parameters (Dredze et al., 2010), or through a domain-specific use of shared parameters.”
Bi-Transferring Deep Neural Networks for Domain Adaptation (2016, methodology): “…alleviating the domain discrepancy: instance-weight adaptation (Huang et al., 2007; Jiang and Zhai, 2007; Li and Zong, 2008; Mansour et al., 2009; Dredze et al., 2010; Chen et al., 2011b; Chen et al., 2011a; Chen et al., 2012; Li et al., 2013; Xia et al., 2013a) and feature representation…”
Towards Universal Object Detection by Domain Attention (2019, background): “Multi-domain learning (MDL) [20, 35, 24, 59, 19, 5] improves on this, by sharing parameters across various domains, and adding small domain-speciﬁc layers.”
Risk bounds for the majority vote: from a PAC-Bayesian analysis to a learning algorithm (2015, methodology): “Indeed, the variance of the margin distribution is controlled by Breiman (2001) for producing random forests, by Dredze et al. (2010) in the transfer learning setting, and by Shen and Li (2010) in the Boosting setting.”
Multi-Domain Learning: When Do Domains Matter? (2012, methodology): “While some earlier research has included a separate one classifier per domain baseline, it almost always performs worse, since splitting the domains provides much less data to each classifier (Dredze et al., 2009).”